Data serves as a mirror to reality, reflecting the intricacies of the world around us through a prism of symbols and numbers. Yet these reflections are not reality itself – they are representations, subject to human error and bias. Like any mirror, the image projected can be distorted – manipulated to serve specific objectives – potentially compromising its authenticity or accuracy.

The universal law of entropy – the natural tendency of systems to drift towards disorder – poses an ongoing threat to information integrity. Without robust DataOps practices, entropy leads to the slow drip of errors and biases during data collection and processing, further distorting data's reflection of reality.

Industrial operations and maintenance decision-makers are acutely aware of this challenge, with 78% of 304 respondents to the Verdantix 2023 operational excellence survey seeing data quality management as a significant hurdle. Industrial DataOps, the blend of ‘Data’ and ‘Operations’, orchestrates people, assets, processes and technology to ingest, trace and quality-control a wide variety of data streams. It contextualizes these streams and delivers them as a self-explaining, digestible and trustworthy payload. This enhances the situational awareness not only of data scientists, frontline workers and C-Suite executives – but also of artificial intelligence (AI) systems.

Why AI systems? The Verdantix 2023 operational excellence survey shows two-thirds of the 304 respondents citing a lack of skilled workers as a driving factor for digital transformation in plant operations. This transformation connects data pipelines from asset sensors, control systems and human inputs to powerful anomaly detection, predictive and prescriptive systems. These systems optimize operations, maintenance and processes to boost revenues, reduce costs and contribute to net-zero objectives.

In February 2023, AspenTech unveiled AspenTech DataWorks by integrating its Artificial-Intelligence-of-Things (AIoT) Hub with its Inmation data management software. This platform empowers industrial decision-makers with situational awareness across operations, enabling the integration and contextualization of data from vendor-agnostic systems such as disparate data lakes and historians. By aggregating this information into a single layer, DataWorks provides a unified, coherent view of the operational environment.

Through the provision of ample context – via data visualization, comparison with physics-based modelling, and collaboration with domain expertise – AspenTech ensures that decision-makers and AI systems are aligned with operational goals. As AI systems evolve to ingest, understand – and act on – not just time-series and event data, but also images, audio and natural language, technology leaders at industrial firms must redouble their efforts to establish and maintain a robust data foundation.

Consider a scenario where a newly installed component leads to a malfunctioning coolant pump. A sophisticated AI system, enriched with historical and real-time data, might infer that the component has been fitted incorrectly. By analysing patterns from similar past incidents, considering the component's lifecycle data and the current operational context, the system can recommend corrective actions or alert maintenance personnel. This level of situational awareness transforms predictive maintenance from a reactive to a proactive endeavour.

The confluence of DataOps and AI within the industrial sphere is forging a path towards operational excellence. These instruments of innovation are sketching the contours of a more astute, nimble and resilient industrial landscape. As we navigate the complexities of data and machine intelligence, situational awareness becomes the compass guiding us to smarter, data-driven decisions – for humans and AI alike.