WHEN Aurizon initiated a comprehensive overhaul of asset management in 2020, the aim was to reduce operating costs, improve productivity and ensure higher levels of fleet performance through exceptional maintenance execution and deployment of advanced technology. The company operates a fleet of more than 600 locomotives, maintained at 14 depots, across a 5100km network extending over the entire Australian mainland. With over 30% of the company’s value and costs attributed to physical assets, maintaining a high level of reliability while keeping costs to a minimum is mission-critical.
Aurizon’s asset management team observed that while many challenges and improvement opportunities are similar across a broad range of industries, the rail sector is slightly different, with low margins and geographically diverse rolling stock and maintenance facilities magnifying maintenance and system inefficiencies.
To provide a bespoke solution for the rail industry, the team developed the Keystone programme that integrates modern maintenance approaches and technology into Aurizon’s established processes.
Keystone systemically and rigorously addresses waste, maintenance practices, system and assets defects, asset reliability and asset maintenance strategies, focusing on:
- data and quality of asset management systems
- core asset management processes
- systems and tools for executing maintenance, and
- developing staff capability.
The Keystone framework is not unique and is consistent with the asset management frameworks used across many industries and for many years. But unlike traditional asset management programmes, Keystone takes a holistic view of the organisation’s asset management needs, including integrating technology seamlessly into day-to-day maintenance processes. Notably, it includes monitoring and intervening in the execution of maintenance at depots.
At the core of Keystone is IronMan, an intelligent, AI-driven asset management system, designed to work seamlessly with an organisation’s existing maintenance management systems and data sources. The tool collects, analyses, and transforms large quantities of data from disparate systems into actionable insights. It uses intelligent algorithms to anticipate and predict potential asset failures, identify trends, and provide recommendations that inform decision-making. The objective is to equip each person involved in maintenance decisions with AI, making them 10-20 times more capable of delivering value.
IronMan was developed by Ox Mountain (OXMT), a British company with a background in Formula 1 motorsport, aerospace, mining and rolling stock maintenance. Working with machine learning and software engineers from the University of Oxford, OXMT created IronMan software a decade ago and has been developing it ever since.
Aurizon elected to trial IronMan after observing great success in the mining sector. After a year of value extraction, the realised and potential value of IronMan convinced Aurizon to invest in OXMT, and it now owns over 70% of the company.
Ironman
One of the key features of IronMan is its ability to access and integrate with multiple systems simultaneously, and handle data that is incomplete or of variable quality. IronMan aggregates and standardises this data and displays it in a form that enables rapid analysis and decision-making. It delivers information and insights in a logical format that is readily understandable by all maintenance staff, from the depot floor to management level.
IronMan takes the mystery, frustration, and potential confusion out of interpreting big data. Using the algorithms and AI, IronMan turns data into information, information into insights, and the insights promote intelligent maintenance decisions.
A benefit for employees using and engaging with IronMan is that it clarifies for them which elements of the data are important for future asset management, reliability and maintenance improvement, and which elements are of little or no importance.
The development of Keystone and integration of IronMan into Aurizon’s operations have delivered profound outcomes for the business. While in hindsight the outputs from IronMan, and outcomes from the case studies (see panels below), seem obvious to anyone experienced in fleet maintenance, in the real world staff are typically too busy to interpret large volumes of data, let alone make decisions based on this data. IronMan has enabled all decision-makers within the depot to use an intelligent system to make informed decisions in real time.
Maintenance practitioners typically create extensive records describing the work they have performed. But if they do not have a system-wide view of how the information is used for asset management, they may miss key pieces of information, such as the nature of the fault that caused a diesel engine to fail, whether an oil leak, power assembly or valve issue.
Key lessons from the Keystone programme are:
- data quality and integration are crucial: one of the most significant hurdles was dealing with the variable quality of data across Aurizon. This variability is not unique to the business, and while the temptation was there to correct the data, that would be a massive and unrealistic task. Companies seeking to replicate the success of Keystone must either have complete data or a mechanism to compensate for incomplete data as IronMan does
- leadership support is key to success: the buy-in from senior leaders was instrumental in overcoming resistance to change and ensuring that the necessary resources were allocated. The commitment from the top down was crucial in driving adoption throughout the organisation
- employee engagement drives adoption: in parallel, Aurizon ensured that the maintenance teams at all levels were trained, motivated, and engaged with the programme. Providing clear understanding and intent helped foster a culture of engagement. Aurizon focused on “doing it with them,” rather than “doing it to them”
- flexibility and scalability are vital: Keystone has adapted to the dynamics of the business, ensuring it remains relevant as maintenance strategies mature and Aurizon’s needs evolve.
New standards
Keystone has demonstrated that integrating advanced technology such as machine learning into traditional asset management practices can significantly reduce costs, improve reliability, and enhance operational efficiency. As a result, Keystone could become a model for other rail operators, especially those managing large fleets and extensive infrastructure.
While Aurizon’s approach is not unique, it is complex, requires cultural change in the business, and demands a robust and powerful machine-learning tool such as IronMan.
Aurizon’s goal is to continue refining and expanding Keystone to ensure that it remains relevant and able to meet future challenges. As the industry moves towards greater automation and further integration of AI, Aurizon will be exploring new ways to incorporate emerging technology into its asset management practices.
Keystone has been transformative for Aurizon. Case studies have demonstrated how it has brought tangible operational benefits, including cost savings and improved asset utilisation. Additionally, Keystone has demonstrated a new standard for how advanced technology can be integrated into business processes, reducing the manual and variable aspects of asset management planning and maintenance execution. It also demonstrates how complex AI tools can be successfully and sustainably used and interpreted at all levels of the business.
The success of Keystone is largely a function of the strong support and leadership from the top levels of the organisation. Led by the CEO, the initiative was recognised as a critical component of Aurizon’s growth strategy. Senior leaders at every depot championed and supported the programme, helping to foster a culture of engagement and ensuring alignment between technology, operations and organisational goals. This drove rapid adoption across all levels of the business.
Lessons learned from Keystone will continue to guide Aurizon’s approach to asset management, ensuring technology is integral to maintenance execution, and providing a template for further improving the business.
Case study: fleet management
THE legacy maintenance model for the Progress Rail GT42 locomotive fleet deployed by Aurizon was built around conventional, manufacturer-prescribed, time-based overhauls. This resulted in the premature replacement of components, wasting up to 30% of their remaining useful life. Time-based scheduling also failed to account for the difference in the distance covered by each locomotive, leading to over-maintenance of low-use units and under-maintenance of high-use assets.
To address these systemic inefficiencies, Aurizon adopted a usage-based maintenance strategy powered by IronMan and SAP. This shift enabled component-level planning based on real-world usage data, including MWh produced and kilometres travelled. IronMan analytics identified failure trends and optimised maintenance intervals. Component replacement was scheduled to optimise useful life and executed in-house at Aurizon depots, reducing reliance on external maintenance contractors. This approach not only minimised downtime, but also cut preventative maintenance costs by approximately 25%.
Fleet availability improved significantly, with a measurable reduction in mission-critical failures per locomotive per year (MCFLY). This new strategy resulted in significant cost savings for the Aurizon fleet. By redistributing maintenance capacity and aligning intervention with actual asset usage, the organisation not only overcame its legacy challenges but also set a new benchmark for locomotive fleet management in the rail industry.
Case study: defect elimination
AURIZON’s nationwide depots, fragmented systems and inconsistent defect elimination (DE) practices were quietly eroding operational performance. Investigation of recurring reliability issues took place in silos, with the staff at each depot using their own tools, workflows, and standards. The result was disconnected data, inconsistent root cause analysis, and a lack of ownership that allowed systemic issues to persist. Valuable insights were not shared and opportunities for cross-depot learning and continuous improvement were routinely missed.
To break down these silos, Aurizon, with the support of OXMT, enhanced the IronMan platform with a purpose-built DE module. This introduced automated investigation triggers, a structured and standardised workflow, and integrated action tracking into a centralised, searchable repository. Investigation is now initiated based on real-time business rules, guided through evidence-based methodologies, and tracked to closure with full accountability. The result is a seamless, organisation-wide approach to DE that embeds best practice into daily operations and fosters a culture of transparency and collaboration.
The DE module has driven a 10% reduction in annual work order spend related to rolling stock defects, while improving process consistency across all depots. Real-time visibility of investigation and corrective action has strengthened accountability and accelerated resolution timelines. Most importantly, the unified system has unlocked cross-depot collaboration, enabling teams to share lessons learned and implement systemic fixes faster than ever before.
*James Petty is head of asset management at Aurizon, where Jared Chiesa is manager, integration, strategy and reliability. This article is based on a paper presented at the International Heavy Haul Association (IHHA) 2025 conference held in Colorado Springs, United States, in November 2025.
