Perfekt partnership and the ‘AI ready mine’: A conversation with Dan Roitman

Leading Australian information technology systems integrator and managed services company, Perfekt is sharing space on the IMARC 2026 exhibition floor with partners, US tech giants Cisco Systems and Everpure, and presenting a host of experts to talk to visitors about “powering the AI-ready mine”.

Perfekt chief technology officer Dan Roitman, who has been with the company for more than 18 years, leads a team specialising in data analytics, big data, IoT and ERP implementation and integration.

An architect of the company’s Asset Information Management and Analytics (AIMA) enterprise platform, Roitman says the Perfekt, Cisco, Everpure and AIMA combination is working with mining organisations aiming to move from disconnected data and isolated AI pilots to practical, operational AI at scale.

Mining Beacon editor Richard Roberts caught up with Roitman ahead of IMARC 2026.

Richard Roberts: What is an “AI-ready mine”?

Dan Roitman: An AI-ready mine is one that has the infrastructure, data, connectivity and technology needed to apply AI across its operations. Not just run isolated pilots but deliver measurable improvements at scale. It means connecting data from equipment, sensors, cameras and operational systems, processing it locally at the edge or at remote sites where real-time decisions are needed and using AI to predict equipment failures, detect safety and operational risks, track materials and optimise production.

With AIMA, combined with Cisco compute and secure networking and Everpure high-performance storage, we help mining companies build this foundation. They can start small with a targeted pilot using, for example, AI and computer vision to measure material moved by loaders. We train and refine the AI model using site-specific data, validate the results and then scale the solution across all loaders at that site before expanding to other sites.

This approach allows mining companies to demonstrate value before committing to a wider rollout, while building a reusable foundation for additional AI applications.

Ultimately, an AI-ready mine is one where data is transformed into timely, actionable intelligence that helps people make better decisions, improve productivity, reduce downtime and operate more safely.

Richard Roberts: Perfekt was founded in 1998. What was the company’s first exposure to mining/metals and how did that evolve?

Dan Roitman: Many members of our team have extensive experience in the mining sector, including developing and implementing technologies across IT and operational technology (OT). I started my career with SAP, working on projects for major organisations such as BHP and Shell, which gave me a strong foundation in enterprise technology and asset-intensive industries.

Over the past 15 years Perfekt has built a significant presence in Perth and Queensland, working with mining and energy customers to address increasingly complex technology and operational challenges. Initially, much of our work focused on solving IT challenges and helping clients manage the growing volumes of data generated by their operations. Data was being collected from trucks, excavators, conveyors and other equipment, but customers often struggled to turn it into meaningful operational insights. They needed better reporting, forecasting and visibility into what was happening across their assets.

As the industry’s needs evolved so did our work. We moved beyond traditional reporting and analytics into predictive maintenance, equipment fault and anomaly detection, remaining-useful-life forecasting, digital twins, process optimisation and computer vision for applications such as material tracking.

This evolution led to the creation of AIMA. We recognised the need for a unified Industrial AI Platform that could bring together different types of operational data, apply the most appropriate AI models and analytics, and translate the results into actionable insights and operational outcomes. Rather than deploying isolated solutions for individual problems, AIMA provides a platform for addressing multiple use cases and scaling successful applications across assets and sites.

Today, our focus is on translating these technologies into tangible operational outcomes. For example, specialised cameras and computer vision can detect foreign objects on conveyors before they cause major equipment damage or unplanned outages, with the potential to reduce related downtime by up to 30%. AI-based measurement can estimate the volume of material moved by loaders and track material accumulating in stockpiles, helping improve material handling and production, with potential gains of around 15%. We can also optimise equipment allocation and material movement routes to improve fleet utilisation and increase overall output, with potential improvements of 5–10%. These are examples of the outcomes we target, with actual results depending on the operation, baseline performance and specific use case.

We’re bringing these capabilities together through AIMA, connecting operational data, video analytics, AI models and optimisation tools in one platform. This helps mining companies move beyond simply understanding what has happened to predicting what is likely to happen, identifying opportunities for improvement, and recommending or enabling actions that deliver measurable operational and financial benefits.

Richard Roberts: When was AIMA commercialised and when was it first activated in the mining/metals space?

Dan Roitman: AIMA was commercialised six years ago. Our first major project was proof of concept for BIS industries and followed by major project with Bradken.

“Over the past 12 months we’ve significantly expanded both the capabilities of AIMA and the range of mining use cases we can address”

Richard Roberts: How prominent a platform is it in the mining/metals vertical today?

Dan Roitman: AIMA is growing rapidly in the mining and metals sector. We’re seeing increasing interest as mining companies recognise the tangible benefits AI can deliver in improving productivity, reducing downtime, optimising equipment utilisation and enhancing operational visibility. As customers see the value of applying AI to real operational challenges, we expect adoption to accelerate further, with opportunities to scale successful solutions across multiple assets and sites.

Over the past 12 months we’ve significantly expanded both the capabilities of AIMA and the range of mining use cases we can address. Our work has progressed across computer vision for equipment and material tracking, AI-based material measurement, predictive analytics and optimisation of equipment allocation and material movement.

We’ve also advanced our approach to deploying AI directly at remote mine sites and edge devices while supporting centralised AI capabilities across multiple sites. Another important development is the integration of generative and agentic AI, allowing users to interact with operational data through natural language, investigate issues and obtain actionable recommendations.

These developments reflect our broader focus: moving beyond individual AI applications towards a unified industrial AI platform that helps mining companies start with targeted use cases, demonstrate measurable value and scale successful solutions across their operations.

Richard Roberts: How much mining/metals-specific development has gone into the platform and how has it evolved a) to meet changing industry needs, and b) as core sensor, connectivity, computing, etc technologies have evolved?

Dan Roitman: AIMA has evolved through years of working with mining and energy customers and understanding the realities of operating in asset-intensive environments. Significant development has focused on mining-specific requirements, from integrating data across equipment and operational systems to applying predictive AI, computer vision and optimisation to challenges such as equipment reliability, material tracking and production efficiency.

The platform has evolved alongside advances in sensors, specialised cameras, LiDAR, industrial connectivity and computing. A key development is the ability to run AI directly on edge devices and at remote mine sites, enabling real-time analysis close to the equipment, faster responses and reduced reliance on continuous connectivity to a central data centre. This is particularly valuable in remote operations, where bandwidth and connectivity can be limited.

At the same time, advances in GPUs and AI models have expanded our capabilities across forecasting, digital twins, computer vision, optimisation, generative AI and agentic AI. AIMA brings these technologies together in one platform, supporting AI at the edge, across remote sites and centrally, so mining companies can scale successful solutions across their operations.

Richard Roberts: How long has Perfekt been working with Cisco and Everpure and to what extent has the alliance focused on heavy industry solution and market development?

Dan Roitman: Perfekt has worked with Cisco for approximately 18 years and Everpure for around 12 years, as an industry partner and systems integrator.

Our collaboration has increasingly focused on bringing enterprise-grade AI capabilities to asset-intensive industries such as mining and energy. Perfekt provides AIMA, our Industrial AI Platform, while Cisco contributes AI-ready compute and infrastructure through solutions such as Cisco AI Pod and Unified Edge, and Everpure provides high-performance, AI-ready data storage.

Together, we combine software, compute and storage into an integrated platform designed for demanding industrial environments. This is particularly important for on-premises AI deployments at remote mine sites, where real-time processing, data security and reliable local operations are essential, as well as for federated and sovereign AI architectures that connect multiple sites while maintaining control over operational data.

Our focus is to turn this technology partnership into practical industry solutions, helping mining customers deploy AI closer to their operations and scale successful applications across sites.

“Most importantly, we focus on outcomes rather than technology alone”

Richard Roberts: Why is this partnership better/best equipped and positioned to support an “AI-ready mine” in a market that today has a lot of attention from a lot of different entities? A lot more attention than, say, five years ago.

Dan Roitman: What differentiates our partnership is that we bring together three critical layers needed to make an AI-ready mine a reality: AIMA provides the industrial AI software and intelligence, Cisco delivers the compute and networking infrastructure and Everpure provides the high-performance storage needed to manage and process large volumes of operational data.

Many organisations can offer individual AI models or infrastructure but mining companies need solutions that work reliably in demanding, remote environments and address real operational challenges. Our combined approach supports AI at the edge, processing data directly at remote mine sites, as well as centralised and federated AI across multiple operations. This enables real-time decision-making while helping maintain data security, operational resilience and control over where data is processed and stored.

Most importantly, we focus on outcomes rather than technology alone. By combining our capabilities, we help mining companies move from isolated AI pilots to a scalable AI platform that supports multiple use cases—from predictive maintenance and computer vision to material tracking and production optimisation. As interest in AI accelerates, our partnership gives customers a practical pathway to build an AI-ready mine, starting with high-value applications and expanding as measurable benefits emerge.