How it works: Document capture for trucking
It’s no secret that paperwork is the enemy of efficiency.
In the trucking industry, however, this problem is particularly pronounced for drivers asked to document deliveries on the road, as well as the home offices left to put the puzzle pieces together for an entire fleet.
Day to day, the red tape often unravels in a “3 way matching process,” where billing departments get bogged down in manually matching transaction information in purchase orders, invoices and proof of delivery receipts — a time-consuming prospect that also adds verification labor costs to each delivery. Ernst and Young, for instance, estimates that it costs around $10 to process each and every freight bill.
But the real reason matching records is so costly is because it’s a fundamental technological mismatch; proofs of delivery are still a physical piece of paper, while purchase orders and invoices can be transmitted in an electronic format and compared by a computer.
Solving the paperwork problem
Until now, record-matching has required manual comparisons to be made by employees. But what if a machine could extract text from documents and match the relevant information automatically? That’s the key to our workflow management tools at LoadDocs.
Using our customers’ templates and documents — freight bills, BOL/PODs, trip sheets, lumper receipts, etc. — we train our algorithm to identify the most relevant fields and values in different documents. By identifying the different types of paperwork, and then the content within those documents, we save our customers a lot of time in two ways:
- Their billing team no longer has to classify or index all the different documents collected from their drivers or owner operators.
- We automate the entering of data in their TMS or accounting software. The customer’s team doesn’t have to re-enter all the data from their freight documents.
To attain those benefits, though, requires several important steps.
Step 1: Training the data
Our process starts by working with the customer to build a training data set.
Here’s an example of a Bill of Lading template used in part to train our algorithm, one of several important types of internal documents (BOL/PODs, trip sheets, lumber receipts) that we rely on.
Step 2: Making documents mobile
Once we have a template, the next step is making sure a document’s structure can be recognized by drivers’ phones.
The horsepower for this function comes from our mobile image recognition engine, which can be accessed from anywhere (which is free to try). This technology can still help in adverse conditions, like low lighting or when images are taken with a shaky hand.
Step 3: Making the ID, extracting the text
Since one key challenge with logistics paperwork is isolating and matching key data, it’s essential to correctly identify and extract relevant information.
Based on the annotated fields, the algorithm extracts the data — noting its position on the page, the label of the section, and the specific area where the value of the data lives.
Here’s a visualization of how it works:
Step 4: Data Integration
After identifying the fields and the values that we’re interested in, such as “Load Number” or “Cargo,” we use the information to populate our document management platform, as well as any TMS, ERP or accounting system.
We integrate with McLeod, TMW, 3GTMS, SAP, Oracle, Quickbooks, Synergize, among others
Once the data is in the accounting system, the real magic can happen: An automated version of the 3 way matching process can take place.
The results are staggering. Customers are able automate their invoice process, saving time and money. If discrepancies are found, end users are still flagged after the 3 way match turns up an issue.
A new reservoir of manpower is unlocked in the process. Instead of contending with all incoming bills, accountants can focus on the exceptions and the errors — the stuff they’re really good at.
Intrigued? Learn more about LoadDocs by meeting our team, trying out a free product demo or dropping us a line.
