Smart Manufacturing is built on solid data and machine learning. While forms of machine learning (ML) have played a role in manufacturing for centuries, it is increasingly finding its form in today’s “Industry 4.0” industrial age – digital, connected, and teeming with data – in more complex models such as generative artificial intelligence (AI).

Data is what drives all aspects of Industry 4.0 – cybersecurity, big data analytics, cloud computing, robotics, augmented reality simulation, IoT/IIoT – and AI and ML are no exception. Thus, it is critical that before diving into implementation, M&D businesses take a careful look at the data sets their AI/ML will run on. After all, as every manufacturer knows, your outputs are only as good as your inputs. 

To position your business to make the most of AI, you need to first make sure that the information it will be working from is accurate, clear, and clean – and safe. Read how.  

AI, ML, and common business use cases for manufacturing

As we continue to explore this new age of AI/ML, we are already seeing ways that they can offer tremendous business value when applied correctly, moving data forward toward new insights and capabilities.

• Finance: Many finance teams use retrospective models to gauge performance and metrics. Gen AI is adept at leveraging non-quantitative, narrative data – e.g., information around service delivery quality – offering more context toward a more forward-looking view. Users can “go back in time,” contextualize data, enrich it, and understand it better in order to get a better view of what’s likely to happen next. 
• Inventory management and resilience: This forward-looking capability can help you better plan for supply and demand, not only to reduce planning lift during “business as usual” but also to weather changes related to geopolitical tensions, natural disasters, and other factors. This might include predictive ordering based on demand data and production times, with adjustments for potential supply chain impacts or weather-related delays. AI can also assist with PEST analyses (Political, Economic, Social and Technological impacts) to determine any onshoring/reshoring needs.
• Predictive and prescriptive maintenance: AI models can use information on factors from temperatures to vibration to help anticipate breakdowns, enabling businesses to remove and service machinery prior to failure, on their own time.
• Sales: If your CRM (customer relationship management) system is riddled with multiple customers with the same name, AI can help you identify who’s who. Compare that to a non-AI environment, where salespeople have to track down each person’s data – or may ignore the multiplicity entirely – either of which can be drawn out and detrimental to your operation.

The importance of foundational data 

The common ground between AI and machine learning is that they both center on data – clean, carefully evaluated data. 

AI has advanced to use complex models as it evaluates data and produces insights and predictions – but these models must be trained. For example, you would need to teach your systems about root causes – i.e. bearings or bushings, wires, and other components that can fail – and how to predict changes in yield, costs, downtime, and other critical operational factors. 

Data is what teaches the model to look for certain trends, outliers, and statistical relationships. The better the data, the faster those AI models “learn” and produce measurable accuracy in their predictions over time. So, they must be trained with quality data that does not introduce noise into the model, which could lead to false or nonsensical insights.

So the question becomes, “What structures and policies are in place to collect that data, and support its quality?” 

Collect, prepare, and protect your M&D data for AI use

Determining the business case for AI and data strategy to support it

As we explored earlier this summer, the first step in AI implementation is always strategy. What are you trying to solve for? Where are you looking to gain efficiency/reduce cost? Consider industry examples to understand how your peers are moving forward and what types of business value are being delivered. From there, data is the last piece of the equation:

• Ask yourself, “What systems are involved in what I am trying to accomplish? What KPIs (key performance indicators) are necessary to monitoring my goals or evaluating my questions? What data inputs will support the outputs I’m looking for, and what is the structure to collect the data?”
• Consider machinery on your shop floor to collect the data: Sensors and widgets used to feed an AI engine to generate qualitative and quantitative outcomes. For IoT or IIoT environments, evaluate whether your machinery is instrumented with appropriate sensors and communication to collect sensory and telemetry data.

Back to basics

Businesses should already have high standards in place for any data they work with: accurate, reliable, complete, accessible, auditable, traceable. Those standards do not change when the user is AI instead of human. Maintain best practices around data management and governance: validate sources, check any calculations, and watch for any data that is siloed, outdated, or formatted in a way that’s incompatible with AI applications. Do you have clear, consistent, standard definitions and formats for master data that defines your assets, customers, suppliers, organization, and location? Consider whether there are any industry frameworks or standards that might be a useful reference – or that must be followed to remain competitive or compliant.

Staff for success

As we’ve discussed elsewhere, the human side of AI can’t be left up to chance. Make sure you have the right people with the right skills to understand the data, build models that use it well, and derive data-driven insights and decisions. Train all staff on the importance of data quality and consistency.

Institute guardrails

Make sure you’re protecting your organization against:

• Threats FROM or INVOLVING data: There are limits to both AI’s process of manipulation and the integrity of the underlying data. The takeaway is to “Trust but verify”: Remember that there is a margin for error and verify information where possible.
  • One “check” on your AI – that can also help you discover business insights – is knowing the standard bell curve or other statistical model that you expect from your operational and performance data. What constitutes an outlier (e.g. two standard deviations away from the mean)? And does that signify an issue in data collection – or does it tell you something about processing and other exceptions that lead to scrap, waste, or other costs? Map outliers back to the issues that caused them so you can correct accordingly.
  • When using Generative AI research to identify and source economic and industry data and metrics responses often veer toward bias or hallucination, Using data directly from public GenAI such as OpenAI ChatGPT or Google Gemini to support your analysis process requires vetting of sources. 
• Threats TO data: Because AI systems often process and store sensitive data, they can be enticing targets for cyberattacks. Take care to develop and enforce robust cybersecurity measures to protect your and your clients’ data and AI systems: regular security audits, employee training on cybersecurity best practices, and advanced security protocols. Understand and maintain compliance with data privacy regulations, such as the EU’s General Data Protection Regulation (GDPR) or the California Consumer Privacy Act (CCPA), to avoid legal complications. Make sure that employees know not to upload sensitive data – such as proprietary or customer data – into the AI large language models (LLMs) that exist in the public domain; consider private and/or enterprise versions of these LLMs, which can exist within the four walls of an organization. And another threat to data in an AI world is the misuse of the data in AI models and outcomes: Poor-quality or poorly controlled AI use can make good data do bad things.

Read more insights from our Cybersecurity team on AI risks and guardrails.

Plan to improve

As with any new business capability, the work isn’t over once a system is in place. Look for opportunities to improve data and how it is used. Consider: How long does it take to get the data answers you need, and why? Is it human error producing data errors? Or cumbersome processes? How can AI remedy these issues? Then, once you know your data is good for how you’ll use it, consider: how can you reuse it? Businesses that have quality data and reuse it multiple times – to improve pricing, channel optimization, marketing, and more – are eclipsing their competition.

In conclusion

Industry 4.0 operates in a digital world where everything’s connected, and AI helps to make the most of that interconnectivity. To survive and thrive, your organization must adopt and adapt to Industry 4.0: That’s not just having cutting-edge packers, palletizers, and production lines, but also having mechanisms in place to capture all the data these new technologies make available, and people who know how to use its insights.

Those who have these pieces in place will be able to move much quicker, with fewer production disruptions and with greater agility to adapt to today’s ever-changing market.