AI is no longer the future. It is no longer set to transform business – it has already begun the process. And many organisations are reaping the benefits it can bring to efficiency and productivity.

According to McKinsey, 50% of companies have adopted AI in at least one business function.¹ The impact of this has been largely positive – six out of ten have seen a revenue increase, while four out of ten have used it to decrease their costs.¹

A recent Deloitte survey supports this view, with 83% of manufacturers believing that AI has made – or will make – a practical and visible impact.² More than a quarter say that the technology has already brought value to their companies.²

However, industrial businesses have thus far been slow to adopt AI. While 83% of industrial businesses believe it can help them deliver better results, only 20% have adopted it.³ So what’s holding leaders back? And how can they prepare their organisations for a productive digital future?

For many companies, it can be difficult to make the business case for adoption – despite seven out of ten businesses seeing a revenue increase from AI use in applications like manufacturing.2 Additionally, the mismatch between the perceptions and reality of AI implementation is a barrier to progress and increased investment in the technology.

“Despite the excitement around AI, there are plenty of inflated expectations of what it can do immediately,” says Amy Challen, Global Head of Artificial Intelligence at Shell. “It’s difficult when you’re going into a business and trying to explain that, actually, it’ll take six months to get everything up to speed. Or that frontline processes will need to change to adapt if they want to see real value from the solution.”

An element of this comes from the fact that, while AI is becoming better understood in industrial settings, there’s still a lot of hype. And this can add to the implementation challenges that undermine AI’s potential.

“One huge misconception is that AI is some kind of silver bullet,” explains Challen. “You get this with every technological revolution. People think it’s magic and that you can just plug it in, everything will work perfectly and you’ll save loads of money. But it’s not like that. Like any big organisational change, it takes time and effort – as well as change management – to get to the point where it delivers real value to a business. People have to understand that just having the solution itself isn’t enough; it’s part of a process that drives wider change.”

Another example is the use of AI across different manufacturing processes. Computer vision is transforming quality testing with cameras able to spot flaws that would be invisible to the human eye. Predictive maintenance combines with machine learning to provide real-time visibility into machine conditions to optimise operations. And generative design in the cloud is allowing companies to complete 50,000 days of engineering in one day.⁴

From Challen’s perspective, AI will act as a catalyst for the decarbonisation of industrial businesses. “AI will be hugely important for supply chain optimisation, electrification and grid balancing,” she says. “All of these obviously feed into the ongoing energy transition. And AI’s role in all three is identifying what is possible and then predicting, managing and optimising demands. So, while the solution to the energy transition may be based in chemical and physical sciences, AI can help us get to this solution faster.”

Even with the examples above, it’s easy to see that AI is a complex technology – and certainly not a quick and easy, one-size-fits-all solution.

For Challen, it’s less about trying to stick an AI solution on top of existing systems and more about redesigning processes to take full advantage of the technology. “The comparison I always use is when steam power arrived in the early 19^th Century,” she explains. “Because water mills had needed to be placed ideally along streams and rivers to draw power, the system was built entirely around that. So, it took easily 30 years for the majority of mills to adopt steam power.”

It’s a concept that rings true for AI implementation as well. “Like the mills, businesses still have everything set up for a pre-digital era,” she says. “Systems are laid out in a specific way for a world before AI and it can be challenging to move away from that.”

Digital technology helps us develop new and adjacent services for our customers

• Shell Remote Sense helps operators of large engines understand when issues are about to occur. This is an example where we combine our expertise in AI with our decades of experience in lubricant analyses to provide a unique service to our customers.
• Shell has over 30 years’ experience providing oil and equipment condition monitoring and analyse 750,000 oil samples every year. Good lubricant management is vital for keeping critical machinery healthy and operations running smoothly.
• Remote Sense uses sensors in conjunction with AI to offer real time oil- condition monitoring. This helps customers to reduce unplanned downtime, optimise maintenance and extend equipment life.

¹ McKinsey & Company. “The state of AI in 2020.” McKinsey.com. 17 November, 2020. (accessed 01 November, 2021).

² Deloitte. “Deloitte Survey on AI Adoption in Manufacturing.” Deloitte.com. 2020. (accessed 01 November, 2021).

³ Aspentech. “Industrial AI Accelerates Digital Transformation for Capital-Intensive Industries.” Aspentech.com. (accessed 01 November, 2021).

⁴ Aspentech. “Industrial AI Accelerates Digi

⁵ Microsoft. “FarmBeats: AI, Edge & IoT for Agriculture.” Microsoft.com. (accessed 01 November 2021).