‘Digital twin’ technology has delivered interactive VR versions of everything from Greater Western Sydney to the Great Barrier Reef – but a business automation expert believes its biggest benefits will come as companies virtualise operations and use AI to optimise them.

If you work in a company, your role is probably part of a core business process like customer ordering, manufacturing, human resources, accounts payable, marketing, and others.

Within most companies, these processes have evolved over time – and there is usually extensive duplication as individual employees tweak core processes to suit ever-changing circumstances and customer requirements.

“Processes are the lifeblood of any organisation,” Dr Lars Reinkemeyer, chief evangelist with process mining specialist Celonis, told Information Age, “and the stunning part is that people always think that the processes are working fine.”

That, however, is rarely the truth once process experts get their magnifying glass out: during a previous role at industrial giant Siemens AG, for example, Reinkemeyer found that the customer order process was being handled in more than 900,000 different ways across the organisation.

“People think they have a clear understanding about how a process goes but reality is so much more complex,” Reinkemeyer explained.

“We’re saying ‘let’s not talk about what’s supposed to be, but let’s get insight about what’s really happening.’”

Process mining explained

Companies are spending big to get that insight, which over the past 15 years has turned process mining – in which companies use analytics tools to better understand what’s actually going on across the business and adjust it using business process re-engineering (BPR) – into a core capability that is expected to generate $3.5 billion ($US2.46 billion) this year and $53 billion ($US46.39 billion) by 2032.

This insight can be incredibly valuable: global kiwi fruit provider Zespri, for one, trialled process mining over the past 18 months and has saved $11 million in just its finance department – while GE Healthcare increased its free cash flow by $1.3 billion within one year.

Not every company can expect such good results, however: companies are terrible at process management, which is why Gartner has predicted that 90 per cent of organisations will fail to capitalise on process mining through 2026.

By that point, the firm believes, one-quarter of global enterprises will have invested in process mining platforms “as a first step to creating a digital twin for business operations, paving the way to autonomous business operations” in which generative AI (genAI) systems continuously monitor and refine the way processes work.

Setting up businesses processes for AI

AI’s ability to rapidly review a situation and evaluate large numbers of potential actions has made it extraordinarily effective in everything from playing the game ‘go’ to optimising a programming routine used in millions of applications.

As the approach becomes more widespread, businesses will develop a standard view of processes that can be modelled as digital twins – not necessarily for you to explore using VR goggles like a house design, but for a genAI engine to analyse for efficiency and suggest improvements.

“We’re visualising how your process is happening in reality,” Rinkemeyer explained. “It’s not a process model, but a fully high resolution ‘MRI’ of the actual process that lets you understand and optimise your process.”

“If you have 9,000 process variants, and your biggest inefficiency is the number of manual touches in those processes, we can help you step through those and make production chains more efficient by seeing where the deviations are.”

Such ‘process intelligence’ – an evolution of process mining that Rinkemeyer describes in a new book called Process Intelligence in Action – will help companies embrace innovations like ‘smart manufacturing’ using new techniques like Causal AI, which models cause and effect relationships and lets managers ask questions like ‘What caused this issue in my manufacturing plant and what actions can I take to prevent future issues?’.

AI use jumps

A recent Rockwell Automation survey of 1,500 manufacturers, including 88 in Australia and New Zealand – found 45 per cent are already adopting genAI and Causal AI, with another 34 per cent expecting they’ll be using the technologies within two years.

“Digital twins were previously used for physical objects like wind turbines or building sites, but can now be applied to organisational processes and supply chains,” said Professor Ganna Pogrebna, executive director of the Charles Sturt University Artificial Intelligence and Cyber Futures Institute (AICFI) in Bathurst, who recently co-wrote a Harvard Business Review paper outlining digital twins’ increasing value in strategic decision-making.

Digital twins “are fast, inexpensive, and advanced, providing managers with unprecedented flexibility to experiment with changes before implementing them,” she said, noting AICFI’s recent successes in developing genAI digital twins for agriculture and planned deployments in sectors like defence and mining.

“CEOs and senior executives can now trial their strategic decision-making prior to execution, fundamentally rewriting the rule book on strategy design.”