Read Time: 5 minutesUntil recently, companies used manual methods to process freight invoices. Some still do. This was mainly because the unstructured nature of BOLs and freight invoices didn’t lend themselves to OCR very well. But now, those issues have been solved with the advent of AI and machine learning. Deciphering complex and diverse form structures is easy if you understand the data like a human and can add context. This is what machine learning does.
Even better, it overcomes the challenges inherent in human processing and auditing. Challenges like:
Human Error
People make mistakes. It’s human nature, and that’s why they’re called “human errors.” This is why many companies use the double-key-compare method of capturing their freight invoices. It makes sense. More eyes on the keying lowers the probability that each keyer and the subsequent quality control person will make an error, and it does. But it is also an expensive process. It costs about 2.5 times more to process a single freight invoice using this method, and you still see errors because people aren’t perfect.
Slow Turnaround Times
Every line must be looked at, every field must be understood, and all pertinent data must be captured, and that takes time, especially when employing the tried and true double-key-compare method. You more than double the time it takes to get (mostly) error-free data from a freight invoice or a bill or of lading. This impacts shippers’ ability to capture early payment discounts and especially impacts carriers who rely on quick payment turnarounds. They both rely on quick, accurate freight invoicing and auditing to positively impact their bottom lines. When any part of the process slows down, they feel it.
Employee Downtime
People also need downtime. They have weekends, holidays, sick days, or sometimes they just quit. In order to efficiently ensure freight invoices are processed on time, and with high quality, employers must often staff multiple shifts or pay overtime costs for working late, on weekends, or during holidays, and that is still reliant on an employee’s willingness to do so. It’s even worse when they leave because of the time and expense of the onboarding process when you replace them.
Scalability
The volume of freight invoices fluctuates seasonally, by client, by carrier, or by simple supply and demand. So, projecting and preparing for them can be difficult and problematic, especially when you rely solely on employees. There is rarely a Goldilocks zone when it comes to processing freight invoices. There are usually too many or too few, and companies are usually struggling to keep up, paying employees to do little, or even laying –off during extended periods of downtime like we saw during the 2020 COVID pandemic and supply chain issues.
The Recipe
Add 1/2 Machine Learning
It’s an easy one; the first part is AI Machine learning. As I touched on at the beginning of this blog, Machine Learning is where companies are going to solve the problems listed above, and it does.
No Human Error
Machine Learning doesn’t make mistakes. It does what it’s taught to do and then goes steps further. During its training period it’s shown where it made a mistake, if it did make some and it never repeats them. Instead, as its name implies it learns from them and seeks better ways to avoid them.
Fast Turnaround Times
Machine Learning works thousands of times faster than a person and doesn’t rely on multiple eyes to ensure that it’s doing its job. It just does it. What used to be 48 or 24-hour turnaround times becomes minutes.
No Downtime
Machine Learning works 24 hours a day, seven days a week. It doesn’t take breaks, and after the initial training period, it needs little oversight. Best of all, it won’t quit on you.
Scalability
Machine Learning can scale up or down as the need arises. The only thing that needs to scale is the amount of processing power you allocate to it.
Add 1/2 Outsourcing
The biggest inhibitor for shippers and carriers to establish their own Machine Learning practices in their own companies are:
- Cost of development
- Specialized knowledge to run it
- The infrastructure
Freight Invoice processing and auditing are Back-of-House operations. Does it require effort, saving money, and ensuring the cash stays flowing? Yes, yes, and yes. But it’s not the core business. Its purpose supports that core business. So the effort and expense of creating a Machine Learning solution are often either on the “back-burner” or given a flat-out “no.” And as someone who just spent the better part of this blog preaching about the virtues of AI Machine Learning, I agree. It makes sense.
Why would a company take the time to employ its own AI solution when it can reap the benefits of someone else’s? They might, but they shouldn’t.
Instead, they should utilized an outsourced vendor partner who has made the processing and auditing of freight invoicing their core business. So much so that they have invested in the infrastructure, servers, technology, and knowledge recourses.
In doing so, a company will see not only the benefits that come with the implementation of machine learning (quality data, speed, and faster payments) but also cost savings. Companies that outsource to a partner with AI ML capabilities can see up to a 50% cost saving over in-house processing or as much as a 35% cost saving over outsourcing to a company that does not have ML capabilities.
It takes both parts to complete this recipe.
½ Machine Learning + ½ Outsourcing = The perfect combination of quality, speed, and overall cost savings.
Let me know if you’d like to know more or would like to see proof. iTech can arrange a free proof of concept. For more information, you can fill out our contact sheet, call, or email jason@itechdata.ai
Subscribe to our blog for the latest industry trends
