Reduction of operational risk is the name of the game when it comes to the supply chain. To achieve a reliable yet lean approach, supply chains must be simultaneously agile and resilient. But both qualities have historically depended on “clean” data. And many organizations weren’t prepared when the need for personal protective equipment (PPE) and toilet paper skyrocketed at the onset of the global COVID-19 pandemic last spring. Business leaders couldn’t wait months for a proper data cleanse to meet the volume of requests for immediate supply. The data had to be both available and actionable. The efficiency of the supply chain became imperative as manufacturers struggled to meet shifting demand.

In traditional situations, a data cleanse would have been called for, but to do that manually is time consuming. Even before COVID-19, the process hampered supply chain efficiency, but once the pandemic was in full swing, slow was no longer an option. Traditional data cleansing has outlived its usefulness, and the time has come to move on. It’s time to let artificial intelligence and machine learning cut through the noise produced by Big Data.

Knowing the Whole Business

Take a step back and look at the entirety of the business operation. While there might be a mostly accurate view of the current facility’s supplies, what’s the state of visibility into other connected facilities? Also, what’s known about other suppliers’ inventories, and how quickly they can fulfill production needs?

Perhaps the I.T. department has a tool that can provide some insight to sister plants in other locations, but there’s little way of knowing what your suppliers have at the same time. The only way around that is to bump up materials on hand to be ready, even if they aren’t exactly what’s needed now. But this approach often leads to having too much of one type of inventory and not enough of another. There isn’t enough real-time information to make meaningful sourcing adjustments to meet production needs.

The long-standing fallback to gather all that information is the data cleanse.

Let the Algorithms Dictate

Organizations everywhere have been handcuffed by project-based data cleanses. They’re expensive, time-consuming, and hurt long-term ROI. It’s not a sustainable strategy, and the process typically has to be repeated every few years.

Even before the pandemic, data cleansing slowed down the supply chain process, degrading business operations. Spending half a million dollars on a data cleanse, which is expected to take up to a year, brings little benefit to the organization. Even with the best technology, the same bad processes are repeated.

Enter AI and machine learning. Cost-effective, quick and easy experimentation is the key to innovative business processes. In this case, it means replacing the old data cleanse with machine learning algorithms, thereby cutting down the year-long process to just weeks.

AI is implemented with existing data to make better decisions and create intelligence immediately, without the use of the data-cleanse process. The AI process takes data out of the silo, so the organization can move from one-on-one relationships to a big-picture overview of the entire supply chain network. The algorithms that are generated with machine learning offer a new level of visibility, and open the doors for others within the supply chain network to share valuable data.

Using AI and machine learning opens up several new opportunities for better data use. One is the ability to do multiple tasks with the data at the same time. For example, if a manager says it is time to reduce inventory, you might normally be hesitant to do it because of the risk of not having the inventory when needed. But in this new model, that’s no longer a problem, because you can better optimize working capital tied up in inventory — not just in your facility, but across organizations within your network. This reduces risk, because you now have the intelligence and confidence that you have what you need, when and where you need it.

This paradigm allows for a smoother transition to Industry 4.0 processes. If there’s a sensor problem on one piece of machinery, the supplier is alerted and provides the part. The real-time use of data eliminates long downtime periods.

Dirty and redundant data should no longer be allowed to slow down supply chain operations. Algorithms generated with machine learning will let you make better decisions with your data, to keep the supply chain functioning at all levels.

Paul Noble is founder and Chief Executive Officer of Verusen,