Document Data Extraction: Is OCR or AI the Way Forward for Freight Forwarders?
Despite Hollywood’s insistence that artificial intelligence (AI) will soon rise to threaten mankind’s very existence, no Terminators, creepy red-eyed computers named HAL, or companies named Skynet have emerged… yet.
While it’s true that AI has now moved from the pages of science fiction to the real world, the technology almost completely manifests itself in practical applications designed to augment the human race, not destroy it – especially in terms of modern industry.
The freight forwarding industry is no exception, and AI data extraction is one area in particular where – far from being a threat – the technology can help firms save time, reduce costs, and generate more value from internal teams.
As Business Grows, Invoices Pile Up
The global freight forwarding market has grown massively over the past decade – from EUR 102 billion (USD 109.9 billion) in 2009, to EUR 151.9 billion (USD 163.7 billion) in 2019.
Such growth is hardly surprising – as the “middleman” in the shipping process, freight forwarders, are at the very heart of an industry that, quite literally, keeps the world economy afloat.
(Image: Size of the freight forwarding market worldwide from 2009 to 2019. Image source: statista.com)
Almost every company that trades internationally relies on at least one freight forwarder to serve as the intermediary between transport companies and themselves.
However, not only do freight forwarders facilitate business relationships between global companies, they also deal with the mountains of documentation that those relationships generate.
Invoices, of course, make up a significant proportion of that stack – all of which contain crucial data that must be extracted and entered into relevant records in the freight forwarder’s TMS or ERP.
However, while logistics is big business and has witnessed huge growth over recent years, many fundamental operational processes haven’t changed much during that time. As Billentis highlighted in a report last year, almost 90% of all invoices worldwide are still processed manually – eating up time and resources that cost. Big time.
No Time for Inefficiencies
In light of the current COVID-19 health crisis, which is causing huge disruptions in the global supply chains, this is no time for inefficiencies.
An April 2020 survey of 300 logistics professionals across the world, conducted by Shipping and Freight Resource, found nearly 60% of respondents indicating that their operations were “significantly affected” by the pandemic. A further 25% said they were “moderately affected” and another 14% “somewhat affected”.
(Image source: shippingandfreightresource.com)
Volume decline, lack of capacity, and delays were all highlighted as specific impacts on operations.
The industry is clearly facing – and will continue to face – massive disruption as a result of the crisis. However, another key finding from the survey indicates that COVID-19 could prove to be the catalyst for long-needed tech investment.
Among the changes companies indicated they will make as a result of the outbreak, more than two-thirds (67%) said that they would be investing in technology – the most powerful result from this line of questioning.
(Image source: shippingandfreightresource.com)
The felt need for technology solutions was also echoed by a large number of survey takers in their open-ended responses.
“Even though we are already online in most activities, we still need to enhance our IT capabilities,” said one. Another indicated that their organization will be aiming to “integrate more digitized workflows,” while another said that “manual processes [will] be seriously looked into for alternatives.”
Traditional OCR Vs. AI Data Extraction
What are the alternatives when it comes to document data extraction?
What is the way forward for freight forwarders at a time when existing pains and frustrations of back office process inefficiencies are being amplified?
The choice largely comes down to two options – optical character recognition (OCR), or AI data extraction technology.
OCR has long been the go-to solution for capturing data from invoices and other business documents.
However, the technology unfortunately suffers from some serious limitations.
The problem stems from the fact that in order to put OCR to use as an invoice reader, numerous templates have to be set up that match each vendor invoice so the technology can actually capture information within the data fields. Individual rules then need to be applied to each and every field.
In practice, this means creating potentially hundreds of templates and thousands of rules – and then continuously maintaining them to ensure those rules don’t break and erroneous data is not extracted.
In short, OCR technology cannot be totally automated. There will always be templates to create, rules to be maintained and manual intervention required every step of the way.
With an AI data extraction solution, on the other hand, you can automate the entire process – reducing the need for staff to deal with labor-intensive manual tasks, while speeding up processing times, reducing costs and minimizing errors in the process.
This is achieved with machine learning (ML) technology.
Machine learning teaches computers to learn through experience – without human intervention.
ML tools encounter data, process and analyze it, make decisions, make mistakes, learn from those mistakes and make new decisions while becoming a better and faster tool than before.
ML-powered AI data extraction tools have already been trained in this way on millions of invoices by the time users plug them into their systems. This means there is no need to build invoice templates or define rules. The technology already understands what all the text it reads in a document means.
With a powerful solution, this data can then be automatically pushed into your ERP or TMS system without any further intervention.
AI Data Extraction from Shipamax
At a time when improving efficiencies is more important than ever and the number one priority is keeping customers happy and the business healthy, AI is the answer.
AI is nothing to be fearful of. AI data extraction tools make logistics jobs simpler and more efficient, while returning a key resource to freight forwarders that they are always running out of – time.
Shipamax is a plug-and-play toolkit for back office automation in logistics.
Our solution uses ML-powered AI data extraction to turn unstructured documents into structured data.
Our machine learning solution plugs directly into your existing infrastructure, syncing directly with any email server to process documents as soon as they hit your inbox. Eliminating repetitive manual tasks, Shipamax enables costs savings, time savings, and greater value generation as your team can finally start focusing on the work that matters.