Breakthrough of European car auctions powered by artificial intelligence solution created in Lithuania
Solutions based on artificial intelligence will make the car sales at auctions much more efficient for Exleasingcar both in Lithuania and abroad. This is a win-win solution for all the auction participants. According to the team at Altic, the artificial intelligence model being developed for Exleasingcar is based on the model created by the US scientists who have been awarded the Nobel Prize. The model is aimed at achieving a breakthrough in auctions. “It’s a multi-stage project, where the first steps are made in data sourcing, selection, analysis of correlations, and building insights to further develop and adapt the artificial intelligence based decision-making models,” explains Marijus Strončikas of Altic, outlining first steps of the project.
According to Mr. Strončikas, the analysis of correlations built on Exleasingcar data will develop and train artificial intelligence (AI) algorithms that would predict consumer behaviour, needs and expectations related to the acquisition of a vehicle and perform multi-stage analysis of predicted purchases and cost-effectiveness. Moreover, the AI based engine would help to manage the process of selling the vehicle at the auction all by itself.
‘It will be possible to make the buying and selling process fully automated and autonomous, thus avoiding the human factor,’ he says.
Altic has chosen to build the solution on Microsoft’s Azure cloud platform and other Microsoft tools. These tools enable building and applying artificial intelligence and machine learning models in the business.
‘Fast and accurate processing of transactions and optimal selection of buyers and sellers are tasks that prove to be difficult for any auction operators. With the help of artificial intelligence, we can complete these tasks much more efficiently, which will allow our company to significantly increase sales conversion. The new system will enable the buyer and the seller to capture the essence from abundant data in the shortest possible time and find the most optimal solution for purchase and sale transaction,’ says Jaunius Barkauskas, representative of Exleasingcar.
According to Mr. Barkauskas, the sales process at the online auction is very dynamic – prices, supply, demand, and participants’ expectations all change extremely quickly. Therefore, information and decisions must be fast and accurate. The systems used in the market are often seen to slow down the process and complicate decision-making due to the limited system functionality and human intervention. The traders expect to eliminate all these limitations with the help of AI.
‘Sellers at the auction aim to make quick sales at a fair price, and the buyers look for reliable product at a fair price as well. Except that the fair price is not necessarily the same for both parties, and someone must come to a decision. In many cases, the decision must be quick and free of bias and errors. In complex cases, such a decision can take time because poorly executed transaction can be very expensive and no one wants to fail,’ Mr. Barkauskas explains the decision to employ the help of AI.
According to Mr. Barkauskas, artificial intelligence can significantly improve the auction sales process by application of mathematical models. As a result, the whole process can be significantly faster due to the elimination of human errors and the ability to efficiently process massive amounts of data every day with modern technology solutions.
The AI will also support use cases for vehicle selection and negotiation engines that will enable the grouping of vehicles for more efficient sales. It will aid in the processes of assessing the price of the car for sale and gauging customer interest, initiating price negotiations, materialising the transactions based on the consumer behaviour predictions and previous sales experience.
‘At this stage, we have already completed collection of all the relevant data, its processing, and preparation for the algorithm application. Afterwards, we will move on to the next phase, i.e., building and training the machine learning models,’ says Martynas Arėška of Altic.
According to Mr. Arėška, the development work involving the development of the final AI tool and integration of data into the platform is performed separately and only after confirming the hypotheses and determining the practical accuracy of the artificial intelligence algorithms.
The development and implementation of the AI system at Exleasingcar’s auctions is expected to take approximately another 9 months. A working group of five IT innovators is involved in the development and implementation of this project.
Exleasingcar UAB runs car auctions in 6 countries. The company belongs to Elca Holding group and is a member of the association CARA.
03 05 2021
