In the last two years, we have witnessed some dramatic upheavals in the trade movement at sea. This resulted in a complete halt of factories at the beginning of the global pandemic, the shutdown of ships due to missing crews and in recent months– unprecedented traffic jams in ports. All of these have severely damaged the global supply chains, and most of the goods are stuck in transitions from ships to land operations. As a result, we often see inventory problems, price increases, and of course, uncertainty.
The problem is that the seaports, the gateway to those goods, is a heavy operation, which, as a more traditional industry, has had difficulty making the transition to the digital world. Eran Pereg, CEO and co-founder of the Israeli startup Conbo, said in a conversation with us at Geektime that more than 90 percent of the port terminals in the world have not made the transition to the new concepts of the Digital Twin, which make it possible to make the leap into the new world data-driven operations, "While in most industries the bottleneck is information analysis, in the logistics sites that are important to our economy, the problem arises with the lack of information collection, not only its analysis," he explained.
Yes, there is a small minority of ports that do incorporate new technologies to collect information with the help of hardware products such as RFID systems, LPR systems, and location sensors on vehicles, but still, the available solutions are complex and expensive to implement. In contrast to those solutions, Conbo, which came out of stealth this week, offers ports a system for managing traffic in their territory, while relying on existing cameras, which are used for security purposes, for example.
The intuition that will help improve traffic in the port
Conbo's development makes use of neural networks fed by what the company calls "information fusion", a process that converts visual information from many sources into structured data. The information is processed, analyzed, and distributed to the parties within the ports so that they can generate operational or business insights from it.
At Conbo, the product is called "artificial intuition," and Pereg explains that this intuition is actually several relatively light neural networks whose purpose is to refine the results with resources and partial information, "based on experience, context, and additional information. As mentioned, a mechanism very similar to the way our brain works, it is similar to the role of the thalamus in processing visual information in our brain."
According to Pereg, standard analytics systems for object recognition classify the video or image without association to a specific area of the port or the time when the incident happened. On the other hand, the system they developed takes all the known information about a specific area – which comes from many cameras that see the same area or nearby areas – and "melts" it together with information about external factors such as weather, the angle of the sun derived from the time when the events happened and the location where they happened. All this information helps produce insights based on context and past experience.
"It may sound like a lot of information to process, but we see that processing information connected to context and experience is the right way to get accurate results and save resources. The accompanying information is the information that feeds the intuition. In other words, we will use all the information at our disposal (and not just the video itself) to obtain initial results. We will then activate several additional mechanisms to verify the results and find correlations to other events over time," added Pereg.
Pereg added that in a short time, the system discovers patterns that enable efficient management of the information processing resources, which leads to particularly accurate results at the same time as reducing the need for processing resources and bandwidth. Another advantage of the product is their technology significantly reduces the amount of information that goes to the cloud – which significantly reduces the exposure to cyber incidents, and security breaches, and enables compliance with strict privacy standards such as GDPR.
Pereg founded Conbo together with Reuven Sabi (CTO). The two came to the idea through a difficulty that Pereg knew from the agricultural world – as he needed to monitor the activity in the open partial areas – partly because of agricultural thefts but also to understand what was happening in the field. The two decided to take on this challenge – monitoring complex areas without using significant processing power – and as participants in the TAU Ventures and Shin bet accelerator, they decided to focus on the maritime sector.
Conbo's development has already been deployed in an American port network and the company is working on additional installations in ports in Europe and Asia. Pereg said that currently, Conbo's competitors are companies that require expensive and complex integrations, which are not always possible at the ports. According to him, the significant advantage of Conbo's product is its cheaper entry level of making ports smart, partly because they do not require any particular installation or integration. The company will sell its product in a SaaS model, depending on the port's actual performance: "If a port moves more goods – they pay more. If we failed to help and his activity was reduced – they will pay less.”
To date, the company has raised $1.9 million in a pre-seed round from several senior angel investors and in a seed round led by theDOCK fund. Conbo has eight employees working from its offices located in Tel Aviv.