Big data and machine learning may help reduce freight costs

Transparency18 sponsored article

Collecting data and making sense of it are two very different things. The transportation industry is flooded with data that is not analyzed. Given its fragmentation, collecting and analyzing it at the company level does not prove cost effective. However, in 2018 making sense of your own data is the new reading skill, companies that don’t do it will be at a disadvantage.

There are two main solutions to this problem. The first one is to build an in-house big data and analytics team — an option that comes with its own challenges and costs. It is hard for non-tech focused companies to run an efficient development team, as there are big challenges in communication between all the involved parties. Trucking is hard to change from the outside without industry specific knowledge and expertise.

There is a second solution, however: outsource your data analytics to a third party. This is not a new strategy to the trucking business, since the industry has always been outsourcing maintenance, trailer ownership, trucking management software and even extra freight. This solution allows for greater focus on the business, instead of making sense of data warehouses, statistics, machine learning, blockchain and an ever expanding field of new technology.

One such solution is Kamion.io. Kamion’s platform combines important tools for today’s freight world and integrates them into a unified solution, presenting the user with the most relevant data based on context. Kamion’s dispatch module integrates with existing fleet management solutions, which saves the business from having to change their existing physical fleet management systems, significantly reducing integration costs.

“Machine learning can bridge the gap between the back office and the truck driver, two worlds considered light years apart until now” – Vladimir Atanasov, CTO of Kamion.io

Kamion uses analytics and machine learning to analyze the available data and provide both dispatcher and driver with smart routing and delivery time predictions. Kamion uses deep learning, the same technique used by google to recognize cats in videos, to build a sophisticated model of each individual driver and accurately predict delivery times, delays and driving patterns. The algorithms build a safety profile of the driver and can detect anomalies in their driving style, which can be caused by fatigue or other, hidden factors.

In the last two years Kamion has been tested with different drivers and successfully prevented accidents based on the data observed. To put this in perspective, the trucking company used for the tests, reduced physical damage premiums by 0.5% and liability premiums by 20%.

Kamion’s platform includes a driver assistant add-on, that uses natural language processing to assist drivers and dispatchers with redundant questions. This technology works across different chat platforms, so communication is fast and uninterrupted. Load relevant  information is available 24/7. “What is my pickup number”,” how many miles is this load”, “how much is my paycheck” are all questions that Kamion answers automatically. That saves both time and frustration while keeping drivers informed.

The platform combines all data channels into one pool and applies optimized algorithms on it, to help dispatchers and drivers make better decisions for a given problem. Would a particular driver be able to deliver a load on time? What is the probability of a driver crashing? What is the best route a driver should take, based on weather, load type and individual driver profile?

Kamion is presenting a demo at Trasparency18 in Atlanta, on May 22, if you want to find out more about the product, the team would love to answer your questions in person.

Stay up-to-date with the latest commentary and insights on FreightTech and the impact to the markets by subscribing.