Smaller and younger technology companies have the advantage of not being tethered to decades-old legacy systems. They can quickly deploy new software that creates smooth connections and processes, and tailor them to market changes.

Industry veterans, on the other hand, struggle to keep pace with innovation. They risk spending years and millions of dollars on a software project that, once complete, will send them back to the drawing board.

One such troublesome legacy process is the bill of lading. On average, logistics companies manually enter information from more than 70,000 scanned documents daily. To complicate matters, no two of these bills have the same layout. So how do logistics companies reimagine their bill-entry systems with innovations well beyond their current capabilities?

Say hello to artificial intelligence and machine learning.

The Need for Innovation

Many logistics companies have bill-entry systems that are 30 to 40 years old, and hosted on mainframes. As a result, they’re having to run large data warehouses in support of their services. And when new clients or employees are added to the mix, the inadequacy of those systems quickly becomes evident. This lack of innovation results in a high cost of training new users, system maintenance and support, and access to knowledge capital.

Billing departments manually sift through thousands of paper documents with multiple templates, resulting in many errors in information processing. By innovating the bill-entry process, logistics companies can better support their clients and provide modernized capabilities, including the use of standardized templates. 

For these reasons, logistics companies are increasingly turning to software providers to innovate and create a digital experience in bill of lading management. Following the choice of a partner for that purpose, the discovery process can last anywhere from six to 12 weeks. It’s essential for teams to be onsite to fully understand the processes, learn how the system works, and assess what needs to be changed.

The first month of the initiative involves an assessment of the problem and determination of improvements that are needed. The remaining eight to 12 weeks see the building out of those recommendations.

The discovery process is about assessing what can be optimized and how to do it. For example, if a logistics company wants specific tables for optimal information entry, the software provider needs to come up with a way to automate that and make it more logical. The whole innovation process is classic user-experience work, with the goal of crafting a solution that addresses all major pain points. 

AI and Machine Learning to the Rescue 

AI can help streamline the bill-entry process, including determining when bills contain too much information. To do that, the technology breaks the business problem into small parts that can be individually engineered and tested. At the same time, software teams can collect historical costs and key performance indicators to gauge the project’s success and build the business case for implementing automation. Combining machine learning with human feedback further aids the AI training process. 

Additionally, eye tracking reveals how employees go through bills, and the kinds of information that stick out the most.

Once problems are diagnosed, it becomes easy to determine how the bill of lading process can be improved. A mockup and prototype of the new process can be shared among all parties, allowing third-party development teams to identify simple fixes to the entire billing-entry application.

The automated system can end up saving employees between 10 seconds and two minutes per bill of lading in processing time, adding up to potential savings of nearly $50 million over the course of the year.

The mission of AI software providers today is to come up with a path forward for logistics companies to automate key processes, and operate as efficiently as possible. 

Ted Eichten is general manager and head of industry verticals at Launch by NTT Data.