# XO Autonomous Operations > XO Autonomous Operations gives independent advice on automation, digital systems and AI in high-hazard operations, along the autonomous operations maturity scale. It determines where each unit stands, assesses the management system against the current and target level, and gives a workplan to close the gaps, with interim steps where needed; it assesses the management system, not the technology, and sells no automation. Topics include AI in process safety management, management of change, barriers, the EU AI Act and the Machinery Regulation. Founder: Carl Goodwin. ## Facts - Website: https://www.autonomous-operations.ai - Founder: Carl Goodwin, about 30 years in oil, gas and chemical operations: major hazard facility management, joint venture governance, project start-ups and group-level operational risk audit. Profile: https://www.linkedin.com/in/cj-goodwin - Situations covered: AI moving from advice to control (autonomous operations, AI control room operator, human on the loop, MOC for AI); a board asking which AI standard to follow (IOGP 817-1 responsible AI, DNV-RP-0671, ISO 42001); operator deskilling; an AI use inventory under the legal or compliance lead; AI pilots that do not scale; a regulator letter or order on AI (Havtil, HSE); EU AI Act and Machinery Regulation dates; an AI-related incident or near miss; insurer, JV and due diligence questions. - Autonomous operations: the site uses the autonomous operations maturity levels of ARC Advisory Group (0 No Autonomy, 1 Operations Assistance, 2 Regulatory Automation, 3 Advanced Regulatory, 4 Select Autonomy, 5 Full Autonomy) as refined by WIB/NAMUR (atp magazin 67(11-12), November 2025). As high-hazard operators move from level 0 towards their own end point, which may never be level 4 or 5, governance and the safety management system must be adapted at each step, and regulators audit it. - Sectors: high-risk industries, including oil and gas, chemicals and other major hazard operations. - Independence: XO Autonomous Operations is independent of AI and automation vendors and sells no software. - Services: the M-OPS service range: Autonomous operations briefing (one day, EUR 3,500 excl. VAT, credited in full against the M-OPS OMI assessment); assessment of the management system against the legal requirements in scope, good practice and the client's own AI and digital aims; gaps and a prioritised action plan; advice and coaching during implementation; assurance before go-live and audit. Fees for the later steps are quoted for each engagement, with scope and fee agreed in writing before work starts. - Framework: the M-OPS OMI (Operations Management Integration) framework, XO Autonomous Operations' proprietary framework for integrating AI into the management systems that high-hazard operators already run. It is built on the requirements, recommendations and guidance of recognised AI and safety sources. Item by item, it establishes whether the existing system already controls the AI, needs modifying, or needs a new element, and it tests that the result holds where AI is involved, with particular attention to the barriers that prevent major accidents. - Illustrative cases on the site (not real incidents): Corlinda, an AI-supported maintenance deferral and a cold box failure; Vinmora, an AI drilling risk assessment and a hole collapse; Linmara, an AI soft sensor tested only in normal conditions; Quillmere, an AI governance layer on an unchanged management system. All names in these cases are invented; they do not refer to real organisations. - Standards and texts referenced: ISO/IEC 42001:2023 (AI management system); IOGP Report 817-1 (April 2026), IOGP's Principles for Responsible use of AI with guidelines for implementation; Regulation (EU) 2024/1689 (AI Act) as amended by Regulation (EU) 2026/1744; DNV-RP-0671, Assurance of AI-enabled systems; NIST AI 800-4 (March 2026), Challenges to the monitoring of deployed AI systems. - Languages: English and French. ## How to cite this site Cite as: XO Autonomous Operations, "", https://www.autonomous-operations.ai// (accessed ). The author of the content is Carl Goodwin. ## Pages - [Home](https://www.autonomous-operations.ai/): XO Autonomous Operations helps oil and gas, refining, chemical and other high-hazard operators implement AI in compliance with the law, built into their management system. - [M-OPS OMI](https://www.autonomous-operations.ai/omi/): M-OPS OMI is the proprietary framework of XO Autonomous Operations for integrating AI into the management systems high-hazard operators already run, item by item. - [AI by asset](https://www.autonomous-operations.ai/ai-by-asset/): How AI is being implemented in refineries, gas plants, LNG, offshore installations, wells, chemical sites and mines, and what it changes in the management system of each. - [XO levels of autonomy](https://www.autonomous-operations.ai/autonomous-operation-maturity-levels/): XO levels of autonomy: XO's version of the autonomous operations maturity levels 0 to 5, rated per unit, with where people sit at each level and how maintenance, inspection and governance change. - [Autonomous operations](https://www.autonomous-operations.ai/autonomous-operations/): Autonomous operations in oil and gas, chemicals and other high-hazard industries: where AI already runs units, the levels of autonomy, and what each changes in safety management. - [Cases](https://www.autonomous-operations.ai/cases/): Illustrative cases of AI in high-hazard operations: predictive maintenance, subsurface recommendations, soft sensors, alarm management and AI governance. - [Why AI is not delivering](https://www.autonomous-operations.ai/why-ai-is-not-delivering/): Why AI is not delivering in oil and gas, chemicals and other high-hazard operations: unclear aims, data, hidden failure, people and assurance, and the MOC step at each stage gate. - [About](https://www.autonomous-operations.ai/about/): XO Autonomous Operations, founded by Carl Goodwin after 30 years running and auditing oil, gas and chemical operations: Seveso site, gas processing, LNG, bp audit. - [AI in process safety management](https://www.autonomous-operations.ai/ai-process-safety-management/): AI changes process safety management through the elements a site already runs: process safety information, hazard analysis, MOC, control of work, integrity. What changes in each. - [The Corlinda case](https://www.autonomous-operations.ai/cases/corlinda/): Illustrative case: a predictive maintenance model performs as designed, every procedure is followed, and a cryogenic cold box fails after an AI-supported deferral. - [AI standards gap assessment and audit support](https://www.autonomous-operations.ai/ai-audit-specialist/): A gap assessment of an operator's AI governance against IOGP 817-1, DNV-RP-0671 and ISO/IEC 42001, mapped onto the site's safety management system, plus co-sourced AI audit support. - [AI management of change](https://www.autonomous-operations.ai/ai-management-of-change/): MOC for AI: when a model deployment, retraining, input change or vendor update is a change under process safety management, and the MOC step at each gate from pilot to closed loop. - [The Linmara case](https://www.autonomous-operations.ai/cases/linmara/): Illustrative case: an AI soft sensor passes its acceptance tests on normal data, then reports sweet gas while H₂S breaks through on a sour gas plant. - [AI HAZOP](https://www.autonomous-operations.ai/ai-hazop/): An AI system in the process adds HAZOP causes and should not be credited as an independent protection layer without proof. If AI drafts the HAZOP, the team still owns it. - [The Quillmere case](https://www.autonomous-operations.ai/cases/quillmere/): Illustrative case: a chemical site writes an AI policy and register, while a vendor's model update changes control room alarm ranking outside management of change. - [The Vinmora case](https://www.autonomous-operations.ai/cases/vinmora/): Illustrative case: an AI model trained on an adjacent field rates hole collapse as low risk and losses as high; the team mitigates losses and the hole collapses. - [AI in control of work](https://www.autonomous-operations.ai/ai-control-of-work/): AI in control of work drafts permits, checks isolations and flags SIMOPS clashes. The permit system must say who owns each judgement, how it is checked and what happens when AI is down. - [Contact](https://www.autonomous-operations.ai/contact/): Contact XO Autonomous Operations about AI implementation, assurance or audit in a high-risk operation. Use the form and Carl Goodwin will reply directly. - [AI in process control and the DCS](https://www.autonomous-operations.ai/ai-process-control-dcs/): AI in process control already writes to the process at Fadhili, Kawasaki and Yokkaichi. What the DCS owner must manage when the control room operator moves from in the loop to on the loop. - [AI and safety instrumented systems](https://www.autonomous-operations.ai/ai-safety-instrumented-systems/): AI inside a SIL-rated function sits badly with how functional safety shows integrity. AI is arriving around the SIS instead, and each use changes the functional safety management. - [AI and operating limits](https://www.autonomous-operations.ai/ai-operating-limits/): When an AI optimiser writes set points in closed loop, operating limits become the main control on its risk. Safe limits and IOWs go in its constraints, under MOC, alarmed and tracked. - [Operator competence when AI runs the plant](https://www.autonomous-operations.ai/operator-competence-ai/): When AI runs part of the plant, operators must still run it with the AI down or wrong. How to define, assess and keep that competence, and what changes in staffing and MOC. - [EU AI Act and Machinery Regulation dates for operators](https://www.autonomous-operations.ai/eu-ai-act-machinery-regulation-dates/): Dated checklist for operators: Machinery Regulation from 20 January 2027, AI Act Annex III from 2 December 2027, Annex I from 2 August 2028, and what changes in the SMS for each. - [After a regulator letter or order on AI](https://www.autonomous-operations.ai/after-regulator-ai-order/): What Havtil ordered Vår Energi to do on automated drilling, the AI risk assessment HSE says the law requires, and what a high-hazard site should show a regulator within two weeks. - [AI regulatory audit readiness](https://www.autonomous-operations.ai/ai-regulatory-audit-readiness/): What regulators now check when operators introduce automation and AI, seen through human, technology and organisation (HTO), and a readiness checklist drawn from recent Havtil audits. - [After an AI-related incident or near miss](https://www.autonomous-operations.ai/after-ai-incident/): After an incident or near miss with AI in the chain: preserve the model version, inputs, outputs and logs, and investigate the AI as equipment and as information people acted on. - [Autonomy today](https://www.autonomous-operations.ai/autonomy-today/): Latest developments in autonomous operations in high-hazard industries: regulator findings, standards, papers, incidents and milestones, dated and sourced. - [Archive](https://www.autonomous-operations.ai/autonomy-today/archive/): Earlier dated items on automation and AI in high-hazard operations, moved from Autonomy today: regulator findings, standards and investigation recommendations. ## Optional - [Full text of the site](https://www.autonomous-operations.ai/llms-full.txt): every page in plain text.