• VoxCroft

Machine Augmented Human Analysts: How a Global Mining Engineering Firm Uses AI to Secure Logistics

Using AI can protect the delivery drivers
Front-line workers like delivery drivers need to be protected. Photo by Macau Photo Agency

With people and businesses going remote this year, delivery people have stepped up to get everything from important documents to pizza from point A to point B but the safety of these front-line workers are often neglected. A global engineering and technology company uses defense industry level open-source intelligence to protect their delivery professionals and other mobile staff in South Africa. Having trained intelligence experts at your disposal has only recently become feasible at this scale with the help of machine learning.

In the United States (U.S.), as in many countries across the globe, being a delivery person, mobile sales person, or truck driver is more dangerous than being a rancher, a miner, or even a police officer. Like police officers, delivery professionals are more active when the conditions are more dangerous the COVID-19 pandemic being a case in point. Whether in crime or protest hotpots, or amid pandemics, and even natural disasters, we rely on the professionals to deliver. In London, drivers fear gangs on mopeds that are out to rob them. In Portland during the 2020 protests, delivery people were caught in the action and arrested.

When you are in charge of logistic security, the potential dangers posed to your mobile staff is near infinite this makes keeping track of everything that could endanger or disrupt the staff an overwhelming task.

An engineering and technology company that services the mining industry, with operations worldwide, contracts with VoxCroft to protect their drivers. In South Africa, the company needed a solution to keep their mobile workforce safe from crime, natural disasters, and frequently violent protests. The typical solutions available to them can be split into two categories: human analytics and machine alerts.

Humans are better at making judgments about situations than machines so a tweet with information about a protest that might be flagged by an artificial intelligence (AI) as relevant, could be judged to be inaccurate or even fake news by a human analyst. Humans are great at finding and understanding context, which machines cannot. Unfortunately, human analytics are traditionally slow and expensive.

What humans are good at versus what machines are good at.
There are things humans can still do better than machines.

This expense has driven many organizations to machine intelligence, which can sift through terabytes of data daily at the fraction of the cost of a human analyst. Unfortunately, current AI technology cannot compare with human intelligence, especially in situations where context is paramount such as logistical security. Customers who use these solutions are bombarded with false positives.

Faced with these trade-offs, the company turned to the experts. The VoxCroft team has over 40 years of combined experience building intelligence solutions for the defense industry. Facing budgetary pressures, the industry shifted from human analysts only to machine intelligence before becoming disillusioned the quality was just not good enough, especially when lives are at stake. Not satisfied with available solutions, the VoxCroft team decided to develop its own solution, optimizing a synergy between machine and human intelligence.

Augmenting human intelligence with AI for risk analytics.
The augmented intelligence synergy between human and machine provides better quality at scale.

The solution is called Project Arrow, and in this case study, we use it for logistic security monitoring and alerting. The platform uses our custom machine intelligence to gather data from all over the world, including social media posts, media articles, newspapers, and even broadcast media such as radio. Expert analysts work with the customer to understand their needs and train the AI to start sorting through the data to highlight only what is most critical and valuable. The human analysts assess the highlighted data, and determine what would be relevant to the customer. Time sensitive events are sent as alerts directly to the customer or delivery person/team. Over time, the machine learns from the analysts and from customer feedback and becomes smarter to exact client needs.

Using AI and Machine Learning for Open Source Intelligence
The Project Arrow technology gathers and sorts open-source data, which is then analyzed by experts.

Drivers can be alerted about protests that would block their route, or natural disasters like fires, and even terrorist threats.

Through the Project Arrow dashboard, decision makers can keep abreast of events or trends that could endanger their mobile staff.

AI Generated Risk Events
The Project Arrow dashboard provides a feed of relevant security risk events.

Logistics managers can get updates on current trends at a glance by looking at trends the machine intelligence has selected based on clustering of events over the last seven days. Heat maps can give a quick snapshot of critical events.

AI Generated Risk Events
Project Arrow generates reports, summaries, and analytics to give you an overview of events.

Having expert trained analysts on call means that longer term trends can be identified and analyzed using their tradecraft. Updates on everything from pandemics to terrorism movements to geographical movements of crime waves can be reported on.

VoxCroft sits at the intersection of human and machine intelligence, which means that we can provide solutions at a scale that was not possible prior to machine intelligence, but at the level of quality relevant when lives are at stake. The Project Arrow are decision makers' eyes and ears to help them ensure the safety of their delivery and other mobile employees.

Schedule a call to discuss how VoxCroft can work for you here.