How AI enables responsible document processing automation for transportation companies
Simon Fauvel Simon Fauvel
November 19 7 min

How AI enables responsible document processing automation for transportation companies

Bills of lading (BoL) have long been a fixture of transportation and logistics. They provide details of cargo, transfer ownership from carrier to receiver, document terms of carriage, and act as receipts. Like any business document, they need to be processed – extracting the necessary data for records and other uses. Until now, that has been a time-consuming, error-prone, and manual job.

Robotic processing automation (RPA) solutions, with their promise of rapid automation combined with competitive pricing, have long been an attractive solution for document processing needs to organizations operating in a myriad of industries. However, as many transporters who have adopted these solutions can attest, they often fall short on their promise to fully automate document processing

In this article, we’ll explore why it has proven so difficult to fully automate the processing of transportation documents, like BoLs and proofs of delivery, and why solutions powered by artificial intelligence are proving to be a more effective approach to responsibly automating document processing workflows.

Why document processing automation hasn’t really worked for transporters

Although technological advancements in RPA document processing technology have enabled more than a few industries to streamline this workflow, there are several issues that have prevented transportation businesses from seeing the same benefits when leveraging this technology.

Lack of standardization

It’s well known there are no standard formats for the vast majority of documents commonly used within transportation, such as bills of lading (BoLs), proofs of delivery, purchase orders, etc. While most of them contain similar elements, such as consignee details and package gross weight and measurements, the reality is that each carrier will use their own format and the degree of variability is quite high.

In addition to differing formats, transporters must also contend with a great deal of variation in document quality. Unlike other industries, like banking for example where documents rarely travel to other locations before being processed, it’s not uncommon for transportation documents to be ferried across multiple locations before they are submitted to be processed. Common issues include skewed documents, poor printing, text overlapping other page elements, illegible handwritten notes or amendments, and general wear one would associate with a document that has traveled to several locations and exchanged multiple hands.

Transporters who work - or who have worked - with RPA products to automate their document processing workflows know that these template based solutions are unable to automatically process documents with a high degree of variability. For example, when a BoL that doesn't fit their templates is submitted, the software “breaks” and is unable to process the information contained in the document. This leaves transporters with two options. They can either choose to invest resources in programming a new template for each page layout they encounter (which is a costly and potentially very lengthy process) or they can resort to entering the information manually every time they encounter a document that doesn’t comply with the available templates within the software.

The issue with both these solutions is that they require manual work in itself – the very thing that this technology is supposed to be reducing in the first place. Even systems that already contain a database of all the most common document formats will still falter when faced with one they don’t recognize, or one that’s become unrecognizable due to a poor print or scan.

AI can extract data from any document

Element AI Document Intelligence is able to process BoLs and other highly variable transportation documents with ease because it was designed with flexibility in mind. Built with cutting edge AI technology at its core, our enhanced optical character recognition (OCR) enables organizations within a matter of minutes to easily set up new extraction parameters and rapidly train the system to extract the right information from the appropriate fields.

And whereas setting up and deploying RPA systems can take anywhere from three to six months, EAI Document Intelligence can be deployed in a matter of weeks depending on an organization's data and the number of integrations they require.

Balancing the convenience of automation and the importance of accuracy

Successful automation is determined by two important factors: speed and accuracy. It’s all well and good to process a higher volume of documents in record time, but when accuracy is sacrificed in the name of speed, the consequences of poorly extracted data are felt all too soon. Winning and maintaining customer trust is crucial in today’s competitive transportation market. Business should not suffer from shipment delays caused by inaccurately extracted data. Alternatively, speed is a decisive factor for most customers when selecting a transporter, making it critical for transportation businesses to effectively calibrate the balance between speed and accuracy when deploying document processing software and supervising its performance over time.

EAI Document Intelligence helps transporters strike this delicate balance by offering an alternative AI-powered approach to responsible automation. This approach is broken down in three simple steps: teach, review, and automate.


When first deployed, users configure within minutes document fields and begin training the models to extract the information. This extraction process is dramatically accelerated by a user-friendly machine learning autocomplete system.


Once the models have processed a few documents, they will proactively suggest answers which are supervised and, when necessary, adjusted by users.

The AI models’ recommendations become more accurate as they learn more and more about the task at hand. This is made possible through human enabled continual learning (CL) – AI that builds knowledge incrementally through human feedback and uses the learnings to improve its performance. This innovative technology is what makes responsible automation possible: AI models that can automatically extract specific data from a document rapidly and accurately, because you’ve witnessed them learning and trained them to successfully accomplish a task with a certain degree of accuracy before automating it.


For most, this is the end goal. When users are satisfied by the models’ accuracy level, they can confidently opt to transition to full automation, creating even more time savings and enabling employees to focus on more complex tasks.

Data extraction is just the first step

Fast and accurate data extraction from your transportation documents is only the beginning. Once data is extracted, organizations need to do something with it. In many cases, data extracted from documents requires some validation or reconciliation. For instance, that could be matching the data extracted from a BoLs to the relevant account information in your transportation management system (TMS).

With the global situation as it is today, shipping demand will only continue to increase, as will the paperwork. Element AI is looking for forward-thinking transportation businesses to work with us to further expand and develop AI powered solutions to answer the unique needs of this thriving industry. If you’d like to improve your document processing workflows, empowering your business to increase its throughput and better deliver on customer expectations, let’s talk.

Ready to fast track your transportation document processing?

For transportation businesses managing on a daily basis high volumes of highly variable documents, Element AI Document Intelligence is one of the few solutions available that empowers them to rapidly reduce processing time from day 1. What’s more, it allows them to gradually transition towards responsible and accurate automation, which in the long haul will result in even more savings. Now is the time to get on that road.

If you’d like to learn more about how Element AI can help you to automate document processing for bills of ladings and other highly variable transportation documents, get in touch with the team today.