With the rise of the internet, and e-commerce in particular, the demands on traditional logistics have increased exponentially. Yet the industry has been slow to adopt the technologies needed to meet those demands. In fact, according to Gartner’s 2017 CSCO survey, over three-quarters of chief supply chain officers admit that their digital transformation projects are still not aligned. Fortunately, change is on the horizon.

While most organizations may not have their digital transformation projects aligned, a study from Capgemini Consulting and GT Nexus shows that the move is already underway. The report highlights that 70 percent of supply chain executives have already launched formal digital supply chain transformation efforts within their organizations, presumably in an effort to capture some of the immense benefits that modern technologies have to offer.

But just how big are these benefits exactly? According to analysis from the World Economic Forum, by 2025, digital transformation of the logistics industry could bring $1.5tr of value to logistics players, plus an additional $2.4tr worth of benefits to society through reduced emissions, less traffic congestion and better prices. Among the technologies expected to generate this value, data-driven information services stand out with an $810bn upside potential for the industry, enabling logistics players to use analytics to optimize routes, reduce maintenance costs and improve utilization.

The value is not hypothetical, either; big data analytics has already proved to be a major advantage for a number of logistics companies. The same WEF report explains that demand forecasting based on analytics has already improved predictions of future demand by 10 to 20 percent, trimmed deployment costs by up to 20 percent, and raised fill rates by up to 5 percent.

We have seen these benefits first-hand through our partners. Just recently, for example, a large European logistics provider was able to predict asset demand on an aggregated country level with 98 percent accuracy two weeks in advance, and with 95 percent accuracy six weeks in advance.

Other companies have leveraged data analytics to optimize delivery routes. Dutch startup Routific, for example, has been able to promise 10 hour deliveries across the Netherlands’ 16,033 square miles. And London’s what3words has been able to improve last-mile delivery time for logistics giant Aramex by over 40 percent using three-word addresses.

As these examples demonstrate, it has become increasingly clear that data, coupled with advanced analytics — particularly when harnessing the power of artificial intelligence — will be responsible for pushing the industry forward.

A Data Quality Roadblock

In fact, according to a recent report by McKinsey Global Institute, supply chain and logistics is one of the sectors in which A.I. has the greatest potential to make an impact. The report points to clear use cases for predictive maintenance of equipment, yield optimization, procurement and spend analytics, and inventory and parts optimization.

A separate report released by DHL and IBM dives further into detail about the various use cases for A.I. in the logistics industry, which Kate Patrick sums up in her article for Supply Chain Dive: “The most clear use case for A.I. in supply chains is harnessing all the data from the supply chain, analyzing it, identifying patterns and providing insight to every link of the supply chain.”

And it’s not just the use of data with emerging A.I. technology that has the potential to turn the logistics industry on its head. Logistics data can also bring major benefits to the industry when implemented with blockchain technology. Outlined in a report by DHL and Accenture, blockchain in logistics has the potential to effect major improvements in efficiency and transparency across the entire supply chain.

However, for the industry to realize the benefits of A.I., blockchain and other technologies, it must overcome some major roadblocks — particularly with regard to data quality. As Richard Waters notes in his Financial Times piece about the McKinsey report, “There are plenty of stumbling blocks. The biggest involve data, starting with how to collect, ‘clean’ and label it in a way that makes it useful for training machine-learning systems.” The same issue of data quality goes for effectively leveraging blockchain in the logistics industry.

Currently, the industry is plagued by various data-quality issues that are preventing the effective use of these emerging technologies. A lot of information is not recorded digitally. There are too many shipments for workers to input data manually. Data is inconsistent and recorded with different types of measurements, in different systems. There is no “single version of the truth” giving organizations a consensus view over the entire supply chain. The list goes on.

To this point, a Deloitte study in 2017 found that nearly half of chief procurement officers named data quality as the main barrier to the effective application of digital technology in their organizations, followed by the lack of data integration. On the other hand, availability of data fell toward the bottom of the list of issues, indicating that the quantity of data is there — it’s just a matter of getting it to the point where it can be used.

A.I. Paves the Way

There is a solution. Ideally, logistics companies would agree with their supply chain partners to implement a standardized system for recording and sharing data, which would give them an accurate and complete view of the supply chain. While traditionally logistics companies have been weary of sharing data in fear of losing a competitive advantage, the benefits of doing so would massively outweigh the risks. But until the entire industry gets on board, companies will need an alternative to improve the quality of their data.

Fortunately, A.I. solutions exist to help logistics companies work with their existing datasets. Machine learning technology can flesh out sparse and inconsistent data, preparing it to be used for the more advanced techniques of optimization, predictive maintenance and forecasting described previously. Using just 5 to 10 percent of accurate, complete data, A.I. can systematically parse companies’ historical data about shipments or inventory, and come up with precise deductions to fix inaccurate or missing data. The result is a complete dataset that can provide real business value.

Applying A.I. to clean data in this same way, companies can take their existing data and transform it into something that can be used to reap all the benefits of emerging technologies. Of course, this requires a minimum level of manual work to first build a training dataset and test the A.I. algorithms, but once that’s complete, A.I. can take the reins and add the substance needed to conduct more thorough analysis and optimization.

At the end of the day, technologies like A.I. and blockchain promise to revolutionize the logistics industry. However, their success depends entirely on the quality of the data that logistics companies feed them. And while data quality is currently a major weak point for organizations across the industry, A.I. is offering a valuable solution. The industry is en route to digitization, and with A.I. it will be able to bypass all of the data quality roadblocks along the way.

Anna Shaposhnikova is co-founder and chief commercial officer for Transmetrics.