Today, we are seeing Industrial artificial intelligence (AI) come of age as a business application for the asset-intensive industries, fuelled by commercial drivers, dimensions of readiness and real-world use cases
The focus on developing, embedding, and deploying machine learning (ML) algorithms as fit-for-purpose, domain-specific industrial applications, is bearing fruit as business drivers capable of delivering sustainable value for asset-intensive organisations emerge. Here we look at these drivers and assess what asset-intensive organisations need to do to prepare for the age of industrial AI.
Drivers of adoption
First, industrial organisations will start focusing on how AI can be applied to address domain-specific industrial challenges. This locks in AI-enabled use cases to tangible business outcomes, making the case for widespread industrial AI adoption.
Second, the barrier to AI adoption will be lowered, as a lack of in-house AI expertise among industrial organisations has historically blocked Industrial AI enablement. More organisations will deploy targeted, embedded Industrial AI applications combining data science and AI with purpose-built software and domain expertise. Fit-for-purpose, embedded AI applications will empower users to tackle domain-specific functions with greater accuracy, reliability, and sustainability.
Third, capital-intensive organisations will shift gears from mass data collection to more strategic industrial data management, with specific focuses on data integration, mobility, and accessibility across the business. That opens the door for industrial AI, and the underlying opportunity for AI-enabled technologies that enable these organisations to unlock hidden value in their industrial datasets.
Finally, the biggest driver for industrial AI are the productivity increases for capital-intensive process industries. Industrial AI enables next-generation asset optimisation solutions to be implemented without relying on large-scale data science expertise. This empowers organisations to bring in new levels of safety and productivity in their workplace, creating semi-autonomous and autonomous processes for collecting, aggregating, and conditioning live data, then feeding it into intelligence-rich applications. The results are new insights, continuous operational improvements, and faster, more accurate decision-making.
Five dimensions to be Industrial AI ready
To get themselves ready, companies need an action plan that maps AI to business goals, data objectives and KPIs to weave Industrial AI into a digital transformation strategy. An overarching industrial data strategy is required – companies need accessible, valuable data that can be leveraged constructively by industrial AI. Building a strategy around quality and efficient data flows is paramount too. That means constructing a pipeline of industrial data that enables AI solutions to process different types and amounts of data required by each use case and application, and scaling this across the organisation so that every user and function of the AI gets consistent performance and results.
A future-proof industrial AI infrastructure necessitates the need to lay the groundwork for industrial AI readiness, requiring collaboration across industrial environments. In fact, the software, hardware, architecture, and personnel elements will form the building blocks of the industrial AI infrastructure. And that infrastructure is what empowers organisations to take their industrial AI proof-of-concepts and mature them into tangible solutions that drive ROI. An industrial AI infrastructure needs to accelerate time to market, build operational flexibility and scalability into AI investments and harmonise the AI model lifecycle across all applications.
Roles, skills, and training are critical. Executing industrial AI relies on having the right people in charge. That means making a deliberate effort to cultivate the skills and approaches needed to create and deploy AI-powered initiatives organisation-wide.
Finally, ethical and responsible AI use is predicated on transparency, and transparency involving keeping everyone in the loop: creating clear channels of communication, reliable process documentation and alignment across all stakeholders.
The above is just a guideline but it is worth approaching things with a holistic view that considers the technical, people and processes requirements, ultimately tailored to your own organisation’s definition of success.
Use cases to bring Industrial AI to life
The starting point of any organisational strategy begins with identifying the business problems, corporate objectives, and strategic goals that Industrial AI can solve. Predictive maintenance is one use case for industrial AI, estimated to have made up more than 24% of the total market in 2018, according to IoT Analytics’ Industrial AI Market Report 2020-2025. Predictive maintenance detects deviations from normal behaviour and prescribe detailed actions to mitigate or solve future problems – all with the goal of optimising output and reducing downtime.
The second use case focuses on quality and reliability. Quality shows how well an object performs its primary function, while reliability shows how well the object maintains its original level of quality over time, through various conditions. Both are significant measurements in an industrial setting and industrial AI enables an organisation to achieve a specific, accurate understanding of the two, saving time and money.
Third, process optimisation leverages advanced machine learning methods, including reinforcement learning and deep learning neural networks, to infer intelligence from different data sources, assets, and processes. With this, organisations can easily identify and mitigate inefficiencies, which have a direct impact on productivity – the primary economic driver of any industrial enterprise organisation.
These use cases are a clear starting point for any organisation building out their industrial AI strategy, and hoping to accelerate time to value in turn. Organisations that successfully put in place the five dimensions to be industrial AI ready and then bring the approach to life through any of these cases will most likely soon be reaping the rewards in terms in improved productivity and profitability, greater operational efficiencies and enhanced competitive edge.
About the Author
Adi Pendyal is senior director of market strategy at AspenTech. AspenTech is a global leader in asset optimization software helping the world’s leading industrial companies run their operations more safely, efficiently and reliably – enabling innovation while reducing waste and impact on the environment. AspenTech software accelerates and maximizes value gained from digital transformation initiatives with a holistic approach to the asset lifecycle and supply chain.
Featured image: ©VAlex
