Maria Joyce, Keynote Speech | ISEAM Event, 13 November 2025:

Opening:

Good morning, everyone. Before I begin, I’d like to acknowledge the Traditional Custodians of the land on which we meet, the Turrbal and Jagera people and pay my respects to Elders past, present and emerging. 

I also recognise that BHP globally operates on lands and waters of Aboriginal and Torres Strait Islander peoples and the deep connection they hold to country, culture, and community.

It’s an exciting time in the mining and resources sector and a privilege to speak with you today around what I see as the evolution of asset management in mining. 

We are standing at the intersection of technological advancement and operational necessity. The decisions we make today will shape the future of mining for decades to come. 

Mining has always been a complex and capital-intensive industry. Our operations span vast geographies, involve intricate supply chains, and rely on a diverse array of equipment and systems. Historically, we have managed this complexity through robust enterprise systems, structured maintenance practices, and a skilled workforce. 

But the landscape is changing. The convergence of data, artificial intelligence, and automation is redefining what is possible. The question before us is not whether we will change, but how quickly and effectively we can adapt.
 
At BHP, our purpose is to bring people and resources together to build a better world. To do this, we need a portfolio that is future-fit and positioned to generate value for decades to come – making today’s topic, critical to our business.

This keynote is not a forecast – it is a call to action. A call to reimagine how we operate, how we collaborate as an industry, and how we unlock further value from the assets we steward.

# 1. The Current State of Enterprise Systems

To start I would like to spend some time setting context on the current state of Enterprise Systems. 

Today’s enterprise systems are foundational to our operations. They provide structure, consistency, and a common computing language across our organisations. 

Today and in the future, our enterprise systems will remain critical to providing a ‘single source of the truth’ to solve and realise our business opportunities. They have been instrumental to BHP in what we have achieved to date in maintenance performance, for example we have one instance of work management and another example the world’s largest database on how equipment fails. 

Today’s systems enable us to manage work orders, track asset performance, plan maintenance, and report on outcomes. However, their capabilities remain largely one-directional, and they lack the ability to act on insights autonomously. This limitation constrains our ability to respond dynamically to changing conditions and to optimise performance across the enterprise.

Asset management systems today are largely designed to solve for individual failure modes. They do not yet hold the capability to orchestrate decisions across broader business functions. Interactions between systems – such as linking maintenance insights to capital and people planning or risk management – requires manual and bespoke effort. While these systems have delivered significant value, particularly in acute high-value events, they are not yet equipped to support holistic, enterprise-level optimisation.

For user, their experience is often fragmented and constrained by multiple data pathways, formats, and computing languages – data integration is not always clean. Data is captured in silos, interpreted inconsistently, and used in ways that limit its full potential. Although we are making data-informed decisions, the richness and reliability of data across domains must improve. Our systems today are not yet capable of dynamic orchestration, and their capability does not support seamless integration across the business.

# 2. The Vision for Future Enterprise Systems

The future enterprise system will be intelligent, integrated, and autonomous. It will operate through a network of Agentic AI models, dynamically interacting to optimise decisions within parameters set by centres of subject matter experts. In BHP, these SMEs exist within our Centres of Excellence model – which was established 10 years ago. Under this model, we have established a centrally controlled globally common approach to asset management. This spans maintenance strategies and governance through to life cycle management, so that we can assess and act on opportunities at the macro, portfolio level. Future enterprise systems represent our next natural evolution. 

Future systems will answer common queries – such as “Which truck within a site’s fleet has components in the best condition to extend the truck’s life by 5,000 hours?”. They will also execute decisions, writing back into budgeting, planning, and execution systems. This marks a fundamental shift from insight generation to action orchestration.

Central to this transformation is the abstraction layer. This layer will serve as the connective intelligence between a miner’s ERP and other enterprise-level systems. It will enable seamless interaction between disparate platforms, allowing data to flow freely and decisions to be made in real time. The abstraction layer will not simply be a technical interface – it will be a strategic enabler. It will allow miners to optimise operations based on their unique resource base, market strategy, supply chain configuration, and operational constraints.

The abstraction layer will also play a critical role in simplifying the user experience. Despite the complexity of underlying data pathways, formats, and computing languages, the abstraction layer will present users with a unified, intuitive interface. This will allow frontline workers, engineers, and decision-makers to interact with the system without needing to understand its technical intricacies. The result will be faster adoption, greater efficiency, and more impactful decision-making.

Moreover, the abstraction layer will become a source of competitive advantage. As enterprise system offerings converge and the gap between providers narrows, it is the abstraction layer that will differentiate one miner from another. Its ability to balance multi-variable equations with dynamically changing inputs – while solving for the uniqueness of each firm – will determine the speed and scale of progress.

Achieving excellence in abstraction layer design will not be easy. It will require deep collaboration between miners, OEMs, and technology providers. It will demand rigorous attention to data architecture, system interoperability, and user interface design. But the payoff will be significant: a truly optimised enterprise system that adapts to the miner’s needs, not the other way around.


# 3. The Role of Data

So what about Data? This is in constant focus for my team and data will remain the fuel for this transformation, but its role will evolve. 

Asset health systems will begin to generate data, not just consume it. Controlled data vocabulary, relationships and logic will be critical to ensure AI models make accurate and reliable decisions. We must invest in data stewardship roles, elevate their importance, and ensure our systems are designed to handle data as an enterprise-level asset, not a domain-specific resource.

Equally important is the untapped potential of unstructured data. Today, vast volumes of operational knowledge – ranging from free-text technician notes and even images – remain inaccessible to traditional systems. These data sources are rich with insights but are difficult to process using conventional methods. Future enterprise systems, powered by advanced AI, will be capable of ingesting, interpreting, and learning from unstructured data at scale.

This capability will unlock exponential growth in our knowledge base. It will allow us to identify patterns, failure modes, and optimisation opportunities that were previously hidden. For example, by analysing historical technician comments alongside sensor data, systems will be able to detect early signs of degradation or recurring issues that would otherwise go unnoticed.

The ability to make sense of unstructured data will also democratise expertise. Knowledge that was once confined to experienced individuals will be captured, codified, and made accessible to the entire organisation. This will support faster onboarding, better decision-making, and more consistent execution across the workforce.

In short, unstructured data represents one of the greatest opportunities for value creation in the digital transformation of mining. As we build systems capable of accessing and learning from this data, we will dramatically expand our operational intelligence and accelerate the pace of innovation.

# 4. The Role of Humans

In every foreseeable scenario, the role of the maintainer will remain, but their role will change. Our workforce will partition into two distinct groups. Both equally important.

Technicians will work “on-the-loop” with technology systems to drive and oversee diagnostics, work priority and planning. Front-line trades will focus on component change-outs using robotic tooling and other technology aids. 

Our workforce will evolve into citizen developers, empowered to build and customise tools within the system. Maintainers will become asset managers, equipped with competencies in electrification, automation, and data analytics.

Training academies will be essential to support this transition and ensure our people are future-ready. The human role will be elevated, focusing on oversight, strategic decision-making, and ensuring alignment between AI outputs with business objectives and physical reality. We will need to invest in skilling today’s workforce to properly use these tools, ensuring that the benefits of technology are fully realised.

The design quality of user interfaces will become even more important, regardless of role type. Intuitive design will be key to the rate and extent of value extraction. As systems become more complex, the ability of users to interact with them effectively will determine their success.

# 5. Machine Technical Limits and Design Evolution

Machine performance will change. We will set a new technical limit in performance and reliability. We will set a new cost floor, lowering maintenance costs.

Today, we operate an imperfect maintenance system, constrained by design and execution limitations. Maintenance does not fully restore machines to their original state. Each service introduces a small loss, which accumulates over time and eventually contributes to replacement decisions.

In the future, machines will be modular, battery-electric, and self-diagnosing. They will compute at the edge, plan and schedule their own maintenance, and execute tasks via robotics. Human interaction will be limited to component change-outs. This evolution will reduce downtime, improve reliability, and eliminate variability introduced by human inspection.

The design cycle will accelerate through AI simulation, enabling exponential advancements in machine generations. Machines will become more predictable, with performance variability eliminated except for environmental factors, like weather. 

Supply chains will be redefined by advances in materials engineering and localised 3D printing, enabling faster and more responsive part replacement. Machines will communicate directly with OEM systems to manage spares and parts demand, ensuring supply chain orchestration is optimised in advance of events.

The transition period to future generation machines will be difficult, particularly as we orchestrate all the elements of system, process and people coming together. As we navigate this shift, we need to engage hearts and minds. 

# 6. Maintenance Practices: From Manual to Autonomous

AI agents will handle fault diagnosis, planning, and scheduling, while robots will execute routine tasks. Instructional content, like work instructions, will evolve into active execution assistance, ensuring process efficiency and quality.

Maintenance will no longer be performed in environments unsafe for humans or machines. Instead, it will occur in controlled settings and increasingly outside of the mine gate. Maintenance execution will become increasingly automated, initially through assistance and eventually through full task execution by robotics.

Most importantly, these advances will make our people safer, both through reduced maintenance demand, like general servicing, but also through robotic tooling that removes people from harm’s way and the line of fire.

# 7. Anchoring to First Principles

As we embrace these technological advancements, it is critical that we remain anchored to the first principles of asset management. 

These principles of asset criticality, failure mode understanding, and return on investment – must guide every decision we make. We need to remain grounded in the maintenance fundaments. 

Not all assets warrant predictive analytics or robotic intervention. We must apply rigour in determining which assets belong in each horizon of asset health: protecting, monitoring, diagnosing, predicting, and prescribing.

The cost and complexity of asset health solutions increase as we move through these horizons, but so does the certainty of preventing failure. Therefore, we must be deliberate in our prioritisation and resist the temptation to apply AI indiscriminately across all assets. 

As steward for BHP’s Way of Maintenance, we will not be introducing new technology for technology’s sake, we will not be moving to future where all maintenance is based on predictive analytics – it will be about making the right decisions, at the right time, for the right assets. Deliberate and considered choices that safeguard what we have built as competitive advantage and serving as the springboard to evolve. 

# 8. The Changing Relationship Between Miners and OEMs

Much of what I have spoken about so far has been internally focussed.  I want to now unpack how is the relationship between Miners and OEMs pivoting.  Which is a crucial piece of the puzzle. 

Today’s relationship between miners and OEMs can be transactional and constrained by proprietary data models. OEMs are experts in machine design, but not in operations. Miners excel in operations but lack design insights. This disconnect slows innovation and limits value creation.

The need for change in this relationship is not optional – it is imperative. The current model, where OEMs supply machines and retain control over critical data, is no longer fit for purpose in a world demanding real-time, integrated decision-making. Miners require open access to machine data to enable predictive maintenance, optimise asset health, and orchestrate supply chains. Without this access, the full potential of enterprise systems and AI cannot be realised.

Moreover, the lack of interoperability between OEM platforms and miner enterprise systems creates inefficiencies that scale with complexity. As miners adopt more advanced technologies, the burden of managing multiple proprietary systems becomes unsustainable. This fragmentation leads to duplicated workflows, delayed responses, and increased risk – a current day example, is when asset health alarms are routed through vendor-specific channels rather than integrated into core business processes.

We foresee that the catalysts for change will be commercial processes, such as RFPs, that will demand open formats and integration-ready solutions. Co-development partnerships will emerge, where miners and OEMs collaborate to build bespoke solutions. Data transparency will become a standard, enabling real-time access and shared insights. Technology disruption, as new entrants from outside mining challenge traditional models, will accelerate this shift.

Changing the relationship is also about unlocking mutual value. OEMs stand to benefit from deeper operational insights provided by miners, enabling them to improve machine design and extend service offerings. Miners, in turn, gain access to more responsive, data-rich platforms that support their strategic goals. This shift from transactional to collaborative engagement will foster innovation, reduce lifecycle costs, and improve machine performance across the board.

Ultimately, the transformation of miner-OEM relationships is not just about technology – it is about redefining accountability, transparency, and shared success. It is about building ecosystems where data flows freely, decisions are made collaboratively, and outcomes are optimised for the enterprise, not just the component. This is the future we must build together.

# 9. Near-Term Challenges

While the long-term vision for enterprise systems and asset management is compelling, the path to realising it is neither simple nor guaranteed. The mining industry faces a series of near-term challenges that, if not addressed with urgency and clarity, risk slowing progress and undermining the value of transformation.

One of the most pressing challenges is data and interoperability. The absence of a standardised “plug” for mining data – akin to the universal USB-C connector in consumer technology – means that integration is costly, slow, and prone to error. Without solving for interoperability, the promise of seamless orchestration and AI-driven optimisation will remain constrained.

The industry must contend with commercial model complexity and vendor fragmentation. Many vendor offerings come bundled with proprietary software that does not integrate well with miners’ enterprise systems. This leads to software bloat, duplicated functionality, and inconsistent data flows. The risk is that miners become overwhelmed by a patchwork of disconnected tools, each solving a narrow problem but collectively increasing complexity and cost. In such an environment, the ability to scale innovation and maintain system coherence becomes severely constrained.

Another challenge lies in internal coordination and organisational alignment. As enterprise systems become more dynamic and interconnected, the complexity of internal decision-making increases. Legacy organisation structures and decision frameworks can act as barriers to change. The risk here is not just delay – it is dilution and loss of competitive advantage. 

Technology roadmaps must be owned by executive leadership, not just by IT departments. The role of the executive will need to change, so that miners can navigate the ethics critical to industry perceptions, ensuring that decisions remain auditable and compliant to legislative requirements.

Another risk will be technology investment and debt. The pace of technological change is accelerating, with ERP and platform cycles now measured in months rather than years. This creates a dilemma: invest too early, and risk obsolescence; invest too late and fall behind. Without a clear and adaptive technology roadmap, miners may find themselves locked into systems that no longer serve their needs, incurring high costs to retrofit or replace them. The danger is not just financial – it is strategic. 

Finally, Workforce capability and skills shortages present a major concern. The mining industry is facing increasing difficulty in attracting and retaining skilled talent, particularly in remote locations. At the same time, the demand for new capabilities – such as data science, automation, and systems engineering – is growing rapidly. If this gap is not addressed, miners may find themselves with the tools to transform, but without the people to deploy, manage, and optimise them. This mismatch could stall implementation and erode confidence in digital initiatives.

Taken together, these challenges represent more than operational hurdles – they are strategic risks. Addressing them requires not only technical solutions but also leadership, vision, and a willingness to challenge legacy thinking. The miners who succeed will be those who confront these challenges head-on, build resilience into their transformation strategies, and maintain a clear focus on long-term value creation.


# 10. Long-Term Opportunities and Value

The long-term opportunity for our industry is profound. We have the potential to extract significantly more value from our limited resources – our ore bodies, machines, and people. This optimisation extends beyond immediate operational gains to encompass broader sustainability objectives.

Our long-term measure of success will be our economic and environmental sustainability. 

By improving operational efficiency, reducing waste, and enhancing decision quality, miners will be better positioned to navigate regulatory pressures, environmental expectations, and social licence challenges. 


# 11. Conclusion: A Call to Action

The future of mining is not a distant vision – it is a near-term reality. The tools, technologies, and partnerships required to achieve it are within reach. But realising this future will require bold decisions, strategic investment, and a commitment to collaboration.

As miners, OEMs, and technology providers, we must move beyond traditional models and embrace a new way of working. A way that is data-driven, AI-enabled, and human-empowered. Let us not be constrained by legacy systems or outdated thinking.

It is about building a mining business that is leaner, smarter, and more resilient. A business that is safer and can thrive in volatility, scale with confidence, and deliver enduring value to shareholders, communities, and the broader economy.

This is the future we are building today. A future for the next generation of leaders.

Thank you.