Are you ready for a black swan event?

Analytics are now within reach of mid-market companies, as ERP platforms are embedding advanced analytics into their systems.

The term black swan dates back to the Roman Empire, and it regained popularity after the September 2001 attacks in New York City and Washington, D.C. In essence, these are rare events that are impossible to accurately predict but have an enormous, often global impact.

A recent example is the global pandemic we all experienced, the effects of which are still with us today. Supply chain disruptions, inflation, the labour crunch—all of these challenges that metalworking companies are facing have been amplified since the pandemic.

It’s true that we know some black swans will occur within a certain time frame. Examples include the outbreak of another pandemic in the next century, the San Andreas Fault producing a major earthquake sometime within the next 200 years, and a large extinction-level asteroid impact occurring within the next 200 million years. But these are large timescales, and knowing exactly when they will happen is impossible.

Analytics, AI, and the Mid-Market

How is it possible to prepare for such large-scale events that are not only beyond our ability to accurately pinpoint, but are also so wide-ranging that their effects are also unpredictable?

Part of the answer is data analytics and AI.

Of course, these technologies won’t tell you, “A city-destroying meteor is 75 per cent likely to hit Japan next quarter, which will create a global shortage of microchips.”

But, if that meteor does hit, AI and data analytics crunches your organization’s data along with data from external sources to tell you what the most likely effects will be, both near-term and long-range. AI also allows you to look at what-if scenarios in the case of shortages and predict what other options may be emerging at a global level.

Unfortunately, many mid-market metal manufacturing companies don’t have a digital transformation road map that includes advanced analytics or AI.

This is not because manufacturers are technophobes, but that deploying, maintaining, and operating it is perceived resource-intensive and cost prohibitive. Until recently, digital tools such as these were expensive, complex to deploy, and typically required data analytics and data science specialists that mid-market manufacturers usually do not engage.

Now, however, analytics is within reach of the mid-market, as ERP platforms are embedding advanced analytics into their systems.

Since the ERP is a fundamental financial and operations technology for most manufacturers, it already contains large amounts of clean, relevant data in a format that analytics can consume. Plus, this data is mapped to business processes making it more usable and intelligent. By leveraging these embedded analytics capabilities, metal manufacturers can prepare not only for large-scale black swan events, but for any unexpected business challenge that they deal with day to day: a supplier who fails to make a delivery, a distributor who suddenly goes out of business, a key customer who makes a massive last-minute order, and even global socioeconomic disruptions.

The Four Pillars of AI

AI provides even more powerful and accurate predictions, but unlike embedded analytics in an ERP platform, the perception is that AI is not as easy to deploy. Just to clarify, AI can be extremely difficult to deploy, however it is achievable with a well-defined digital transformation journey that employs a crawl-walk-run approach.

To be completely transparent, though, deployment is just the tip of the iceberg.

Using AI effectively requires complete buy-in across the company, with strong support from leadership and cross-organizational collaboration to build and execute a well-communicated vision, a focused mission, and an organized strategic plan. It’s a cliché in tech to say that a project is a “journey,” but in the case of AI, this is absolutely the case and should also be part of the company’s overall digital transformation. Remember, AI is not a set-and-forget initiative; it is a living and breathing initiative that must adapt as the organization grows and changes.

With analytics tools that support your vision, you are ready for AI. Start your AI journey with a specific objective in mind that, once achieved, can rapidly add value to the business.

Whether that’s enabling multi-sourcing to increase supply chain resiliency, forecasting demand more accurately, or something else entirely, having a strong, valuable use case will make an enormous difference. ChatGPT, for instance, is a very hot technology right now. But as fun as it might be to play with, deploying generative AI without a clear objective for its use is a perfect recipe for failure.

You’ll also need to ensure you have the four pillars of AI in place, which are required to create value for your company.

1. The right people. Your company needs a diverse team with specialized skills, the first of which should be a data scientist or a data analyst. AI requires ongoing training on a lot of relevant, well-formatted data to make predictions that hold up and are free of biases. You’ll also need technologists who can assess the underlying infrastructure needs, which can be significant, along with business process owners who can make sure AI is properly integrated into operations.

2. The right processes. AI must interact and be integrated with business processes and the people who manage them. If the processes aren’t right or if the processes can’t incorporate AI insights, then AI may as well not exist.

3. The right systems. The infrastructure requirements for AI will vary according to the use case, but they can be substantial. If AI is working with very large datasets, you may need high-performance computing (HPC) clusters that include high-end graphics processing unites (GPUs). Data may need to reside in a data lake (a centralized location designed to store, process, and secure large amounts of structured, semi-structured, and unstructured data) and IT will need to make decisions about whether cloud, on-premise, or a hybrid approach strikes the best balance between price, performance, and security.

4. The right data. AI needs a wide array and large amount of relevant data. If you still operate HR, shipping, and other business functions using paper-based processes, these must be digitized before you implement AI; otherwise, AI cannot access any of it. If you have the right people—data scientists, in particular—they can identify the right internal and external data sources to achieve your objective.

Preparing for the next black swan is all about insight, information, and agility. Both AI and analytics can provide rapid, accurate insight into a fast-changing world. And while AI may be more powerful than analytics, it’s also far more complex to deploy and integrate into an organization. Starting with analytics, especially analytics that are already embedded in an ERP platform, is a good first step on the journey towards AI and building a data-driven organization that can pivot properly in the face of unexpected events.

Gavin Verreyne is senior vice-president client success and digital transformation at SYSPRO, www.syspro.com.