Document Automation

Document Process Automation: Why it’s a Game Changer for Freight Forwarders

Data entry and extraction. It’s time-consuming work that requires high concentration and high effort, often from key members of staff. Nonetheless, it is often unavoidable – especially for freight forwarders handling high volumes of documents. 

Or is it?

Research from Drewry Shipping Consultants estimates that container shipping lines deploy some 5,100 containerships worldwide and provide approximately 400 scheduled liner services, most of which sail weekly. They carry about 60% of the goods that are moved internationally by sea, generating some $166 billion in transport revenues globally, along with 1.26 billion freight invoices – all of which must be processed at some level by those hard-working data entry teams. 

And here’s the thing – as highlighted in a Billentis report last year, of all the invoices issued worldwide, a jaw-dropping 90% of them are still processed manually.  

The Costs of Manual Data Entry

The tedium and monotony of manual data entry is not something freight forwarding companies can afford to ignore. For one thing, bored and frustrated workers are not happy workers – and unhappy workers are not productive. In fact, according to a six-month study into happiness and productivity from the University of Oxford’s Saïd Business School, published in October 2019, happy workers are 13% more productive than their dissatisfied counterparts. 

As co-author of the study, Professor De Neve points out, “There seems to be considerable room for improvement in the happiness of employees while they are at work. While this is clearly in the interest of workers themselves, our analysis suggests it is also in the interests of their employers.” 

Indeed, it is – and data entry professionals are feeling this frustration. 

A separate study from last year found that 63% of workers feel that repetitive tasks like the “dull dread of data entry” are robbing them of the opportunity to show their best work. And this can have further consequences. 

When employees are bored, tired, frustrated, or even embittered with their work, mistakes are made – and mistakes in data entry can prove to be hugely expensive. Stories of spreadsheet fails make for great, wince-inducing reading (see ‘12 Small Typos that Caused Big Disasters’) – provided, of course, you don’t find your company’s name on the list.

But freight forwarders don’t need to make a multi-million-dollar blunder to throw money away through their data entry operations. If you’re still relying on manual invoice processing, as 90% of companies seem to be, you’re doing it already.

Take email, for instance – the number 1 form of data exchange in the shipping and freight industry. 

Past research with freight forwarders has shown us that invoice attachments can take at the very least, 10 minutes to extract by a data entry professional – and up to 60 minutes for larger documents. A time consuming task and often erroneous task that often requires fixing (by an agent). 

Doing a few quick sums, this means that if you’re paying your clerks $20 per hour to process a maximum of 6 documents per hour – each document you receive is costing you at least $3.33. 

Multiply this by 90% of Drewry’s figure of 1.26 billion invoices (so, 1.13 billion invoices) and we see that the industry is facing a $3.77 billion bill each year just to handle data extraction and entry from email attachments. 

It’s Time to Change the Game

When relying on manual processes, data entry and extraction is a risky, expensive and soul-destroying game. 

Thankfully, there’s a better way to play it – and that way is with AI-powered document processing automation. 

The Drewry report notes that shippers and freight forwarders have been somewhat slow to adopt automation – and this is costing them. Big time. 

As well as sky-high payroll costs, there is also the cost of customer satisfaction to be considered. 

Manual processing leads to slow turnaround times and high error rates, which cause delays for your customers. Combined, Drewry estimates that process inefficiencies due to low levels of automation and lack of trust are costing the freight industry $34.4 billion annually. 

While this figure of course, takes into account the entire invoice processing lifecycle, it’s clear that by failing to automate the steps that can be automated – such as data entry and extraction – billions of dollars are being left on the table. 

So, how does automated data extraction work and how can it change the game for freight forwarders? 

AI-Powered Document Automation Software  

You may already be using certain technologies such as optical character recognition (OCR) to extract data from invoices and other business documents. But OCR has some serious limitations.

Put simply, OCR recognises text from image-based documents – such as PDFs or scanned invoices. You can read more on OCR vs RPA vs AI in this short blog.

The problem is that OCR-only solutions are only really useful when processing documents of low variability. The reason is that for each invoice a freight forwarder receives, the technology needs to be taught where the relevant data is located on the document so it can extract it. This is done by building templates of the invoices and setting rules so the technology can then collect data fields from specific coordinates within the digital image.

Herein lies the problem.

Freight forwarders receive thousands of different invoices from dozens, if not hundreds of different companies – and no two invoices are structured in the same way, nor necessarily use the same language (“Total”, “Amount Due” and “To Pay” all mean the same thing, for example). 

As such, for every new type of invoice, a new template with new rules must be built and then maintained – an ongoing process that can often eat up as much time as it saves. 

AI-powered data extraction, on the other hand, needs no rules, no templates and no human intervention to extract data from one document and push it into your ERP or TMS. Instead, the best solutions on the market today utilise machine learning (ML) to perform the entire task automatically. 

Machine learning document processing automation tools are trained on millions of invoices to learn all the different patterns, structures and terms they contain. As such, by the time a freight forwarder plugs the solution into their system, the technology already understands what the text reading means. 

In other words, machine learning brings true contextual understanding to invoice data capture – and this has a whole host of benefits. 

For starters, ML-powered solutions that automate both data extraction and entry dramatically reduce the time it takes to process invoices. And as they save time, they also save money. We know that document processing automation can reduce manual data entry costs by more than 80% – from $3.33 per document down to just $0.60. 

And not only do such solutions process documents quicker, they do so with an extremely high level of accuracy – up to 99% accuracy in fact. 

And of course, this accuracy can be relied upon day in and day out. Machines don’t get bored, tired or careless – they “happily” perform the same dull, repetitive tasks around the clock and calendar, without taking breaks and without making mistakes. 

Final Thoughts

Your employees are dying to show you what they’re worth. 70% would welcome the implementation of document processing automation tools to handle data entry tasks, according to the survey. This would make them happier in their work and free them up to focus on more valuable tasks.

ML data extraction and entry technology can make this happen, processing hundreds of invoices in a matter of minutes for a few pennies per document. 

In short, AI-powered automation technology will save time, save money, reduce errors and increase employee happiness (and thereby productivity and retention) – speeding up operations and leading to stronger relationships with better-served customers. 

It is the game-changing technology freight forwarders need. 

Josh BradleyVP Demand Generation
June 2020
9 min read
  • Document Automation
  • Data Extraction
  • OCR & AI
Share this post