By Nicol Ritchie, DataProphet technical writer
The manufacturing leaders of today contend with ever-intensifying pressure to evolve their factories to become digital-first. Manufacturing Lighthouses, for example, are a global proving ground for the idea that intelligently connected plants are not only more market adaptive, efficient, and productive but also have more engaged and collaborative workforces.
Resistance to smart-factory initiatives…it’s only human
Organizational change veteran, Erika Anderson, reminds us that humans are hard-wired to perceive an imposed change to their habitual way of doing things as initially ‘difficult’, ‘costly’, and ‘weird’. Consideration of Anderson’s “Change Arc”, therefore, is a great starting point for the digital transformation journey, a business journey as human-dependent as any other.
With natural resistance factored in, her Change Arc anticipates the questions people typically ask when faced with workplace change—What’s changing for me? Why? What will it look like? The big idea here is that if the change is motivated contextually, this establishes a safe psychological basis for personnel to progress through an adopted change. The progression itself cultivates “change capability”. It is justified in the results by the change being re-envisioned as ‘easy’, ‘rewarding’, and ‘normal’.
The behavioral change arc towards an intelligently connected factory should not be underestimated. This is because, somewhat paradoxically, technology may not be the biggest hurdle to the digital transformation of production. In some cases, the deciding factor in 4IR (Fourth Industrial Revolution) adoption initiatives can be organizational dynamics. Human factors of resistance to implementing data-driven technology apply across the board—from the plant personnel and IT systems owners at whose fingertips advanced analytics optimization solutions are designed to be placed, to the C-suite, whose wholehearted mandate for machine learning (ML) use cases is an essential precondition.
The age of data—overriding the “no change” management myth
Unfortunately, the conventional business wisdom that 70% of organizational change management initiatives fail persists despite this narrative being rigorously challenged. Some of the reasons this study gives for a negative bias towards the idea of successfully changing organizational culture are instructive. Namely, systemic human-behavioral change in the workplace is highly context-dependent and occurs over time. Therefore, measuring change impacts has often been fraught with ambiguities.
However, in the current era, three factors mean the impact of digital-first change in plants is achievable and measurable:
-
A wealth of high-resolution industrial-edge information from machine sensors and other systems connected to the production line.
-
Distributed system technology stacks to capture and connect this data across ISA 95 levels 0-4.
-
An associated sophistication in data science techniques to curate, leverage, and interpret the data.
These technological factors can combine in intelligently connected factory environments to render change impacts objectively measurable. In other words, when data is harvested correctly, it can function as a source of truth disrupting the narrative that organizational change is not quantifiable. This is especially the case when operational technology and human interaction with that technology are interlinked.
Driving a digital-maturity culture with specific value targets
On successful digital transformation, Kate Smaje and Rodney Zemmel, global leaders in McKinsey’s Digital practice, advocate for a value-oriented approach where initial momentum is harnessed via a heavy concentration on one business function. On the other hand, Forbes Councils Member, Ricardo Buranello, echoes their conviction that digital maturity in business is a culture-first phenomenon.
Indeed, for factories to become “smart”, a shift towards an enterprise-wide digital maturity mindset is inevitable. However, this psychological shift is ultimately predicated on plant managers and their teams enjoying first-hand exposure to extracting additional production performance value from new technology at the industrial edge.
Herein lies some excellent news for manufacturers looking to drive new business value by digitally transforming their standard operating procedures to be more data-centric. A singular, smart factory use case can, with some caveats, innately catalyze digital transformation.
The collective digital-maturity mindset in practice
The potential value of industrial data to optimize production is often underappreciated or misunderstood by plant operations personnel. One reason is philosophical. There can be mistrust of the validity of data science as applied to a material process. Data Scientist and Engineer, Jan Combrink points out that deep learning prescriptions, for example, can seem counterintuitive to plant operators:
“An operator’s sense of the process they’re overseeing is technical in the immediate sense; it’s an empirical understanding. On the other hand, a plant engineer has a theoretical sense—far removed from the actual process. Just a very high-level overview of everything modeled in a fine-grained manner, but abstracted. And I think that when a person has an extensive background in engineering, the mathematics of the optimization problem, if explained, makes sense to them. And there is a little bit of a gap there between those two paradigms. The plant operator is like—‘Why would I make these changes?’”
Technically speaking, this conceptual gap is because the optimal operating region AI defines is non-Euclidean. It does not adhere to the rules people intuitively expect.
Echoing Combrink’s first-hand factory experience, international innovation consultant and CEO Mike Walsh categorizes the conceptual paradigm necessary for our age as computational thinking supported by algorithmic leadership. To be sure, plant personnel buying into the underlying validity of data-driven insights into the material flow of manufacturing requires multiple stakeholders to work in a spirit of collaboration.
For a manufacturing process optimization solution powered by prescriptive analytics—plant engineers, plant managers, the AI vendor, and the C-suite need to communicate regularly from an underlying belief in the base technology. It is up to the AI vendor (ideally, with the support of an in-house champion of the project who has data science expertise) to explain the solution and prove its value transparently.
However, this cannot be smoothly achieved without a commercial mandate at an executive level for the AI use case and a subsequent technical mandate from the plant engineers that the plant manager also gets behind.
Successful smart-factory initiatives align multiple stakeholders around changing for value
AI-as-a-Service recognizes the importance of aligning these multiple stakeholders in unique and dynamic plant environments. Unlike a plug-and-play solution, data science specialists and install teams work backward from an end value goal (expressed in terms of ROI). This goal is agreed on between the vendor-partner and the manufacturer, for a specific product or several products. It is usually pegged to a defined acceptable range of improvement in yield over quality within a projected time period (between eight months and a year).
Compliance with AI prescriptions is critical here. This necessitates plant engineers, operators, and IT work with the AI vendor to enact the changes necessary to move the plant to an optimal operating region. This happens at the intersection of the factory’s OT (operational technology) systems and the human-machine interface (HMI). It requires a normalization process during which personnel’s habitual way of doing things goes through subtle modifications that must be integrated into their workflows.
In light of the Change Arc mentioned earlier, this process maps well onto the change-capability trajectory. Against transparent, data-derived value tracking, the results of the digital transformation can be assessed as measurably achieved (i.e. seen to have yielded additional and sufficient commercial and efficiency benefits).
In this event, plant teams and the C-suite can pause to appreciate the fact that the operational behavior for that particular business unit is emblematic of a practical digital maturity mindset shift, from ‘difficult’, ‘costly’, and ‘weird’—to ‘easy’, ‘rewarding’, and ‘normal’
Now, let’s take a step back.
Human-machine data actualization cycles
Along the way, it is essential for plant teams and execs to realize that an algorithmic model of a production line needs to see how the different sub-processes relate to one another. Reverse engineering à la first principles, the fuel for 4IR optimization is the plant’s historical and actual industrial data (both process and quality).
Crucially, therefore, a digital change management process geared to optimize production via deep learning prescriptions begins by ensuring the factory has:
-
Data integrity
-
A comprehensive data picture of a targeted production line
Depending on the factory context, data is collected in various ways: via programmable logic controllers (PLCs), Supervisory Control and Data Acquisition (SCADA), and Manufacturing Execution Systems /Quality Management Systems (MES/QMS). Factory personnel generally have an awareness of different data sources but have often never properly utilized the database themselves.
As Combrink explains, “At the start of every project, there is a handover of a plant’s data. However, in reality, it usually manifests as a disjointed set of spreadsheets and databases. For example, quality measurements of production phenomena arrive that were measured in a lab and written down and have been entered in. However, these data sets tend to be collated and handed over without context, which is to say without the customer having a full picture of the data’s relevance.”
Indeed, some data may have been collected for purposes that are not even known to plant personnel—regulatory requirements, for example, or sensor measurements logged for the purpose of non-defined future use, thus lacking a plan.
This is why an AI-as-a-Service vendor collaborates with the manufacturer to address the factory’s specific data challenges (in terms of data integration and the requisite human processes needed to achieve it) at the outset. An AI-readiness assessment is in fact a data discovery journey that plant personnel embark on with process optimization specialists. The data discovery journey itself adds significant value to the plant’s standard operating procedure. It may do so by revealing critical areas for improvement—in data logging across the production line; refining the data; improving logging resolution; logging of additional parameters, and implementing better product traceability.
In conclusion, AI-as-a-Service served by a committed partnership ecosystem and the right technology can expedite the digital maturity adoption curve. Done right, a singular smart factory use case intelligently connects a plant by collaboratively aligning stakeholders at the nexus of next-generation technology and professional growth. Successful smart factory use cases can then be scaled across production lines and even to the level of the factory fleet.