From Legacy Systems to AI-Driven Innovation

Pioneering Agile Development and Enhancing Efficiency in the Automotive Industry




Audi Used Car Plattform


Technologies & Tools


What we achieved together

Enabling innovation by connecting a modern sales platform with existing systems

As a renowned German car manufacturer, part of Audi's B2B portfolio is selling used cars to their dealership partners. To improve profitability in this context, a legacy solution had to be replaced with a quickly adaptable web platform that allowed room for innovative sales formats and AI-driven pricing. This project is driven by the Audi Software Development Center (SDC) which focuses on individual software products that offer a significant market advantage. The SDC wants to set a new standard for agile software development and thrives for a faster product development, higher quality and an increased depth of value creation in the company.

The project was set up with a product owner from the business department and a technical product manager from Audi SDC who is responsible for the architectural alignment and integration into the IT landscape. Audi chose Comsysto Reply as its partner, providing a cross-functional delivery team of 10 people.

The first step of the challenge was to re-implement the existing functionality - selling used cars to the dealerships in form of live auctions - as a modern web platform. The cross-functional team achieved this by building a Spring-based back end and a React-based front end, and hosting the developed system on AWS using infrastructure as code with CDK.

A short time-to-market was enabled by the deliberate decision against a complete green-field-approach and instead integrating with other systems of Audi's digital landscape using web interfaces (REST/SOAP) (figure 1). As a consequence, the MVP was designed to work seamlessly with the existing Excel-based process of generating auction files out of an Audi host system by providing a corresponding import functionality. This mechanism was later replaced by an even more sophisticated interface.

Figure 1: Overview of the MVP’s system context.

The new user-centrically designed, reactive UI with live push notifications over WebSocket (figure 2) received very positive customer feedback and provides a much more pleasant buying experience than the legacy solution.

Figure 2: Screenshot of the platform during a live auction.

With this foundation in place, new sales formats were explored using an iterative and lean approach. One milestone was marked by the introduction of a marketplace that enabled selling cars outside of auctions and their fixed time schedule, and with new pricing strategies. By now, the marketplace accounts for around 30% of all sold cars.

By providing the platform's sales data to one of Audi's central data warehouses, it could be used to predict used car prices using machine learning (figure 3). Using these AI-predicted prices, the sell latency of cars could be decreased by roughly 50% due to the prices being closer to their real market value, thus reducing costs for Audi and effectively increasing profitability.

Figure 3: Overview of the system context after the integration with an AI-driven pricing service and the replacement of error-prone manual process steps.

Another major benefit was that complex and error-prone manual steps of the old sales process were left behind while other steps were automised, making the whole process simpler and much more efficient (shown by the absence of the manual process step when comparing figures 1 and 3). Combined with the fact that the new platform provides the much needed development flexibility for further innovation, the project has already positive implications on Audi´s used car business.

All in all the project can be considered a huge success. Since going live in August of 2021, a total of around 29,000 cars were sold on the platform while routinely handling live auctions with an average sales volume of almost 8 million Euros. Accordingly, the Used Car Platform is continuously developed further, and with quite an impactful roadmap ahead.

I had the opportunity to collaborate with Comsysto Reply on the further development of the Used Car Platform. A particular challenge in the project was to integrate the modern cloud application into an existing system landscape with legacy systems, thus achieving a rapid go-live. Comsysto Reply significantly contributed with outstanding expertise to sustainable solutions. The collaboration was always partnership-driven, which is why I would recommend Comsysto Reply at any time.

(translated from German)

Fabian Strathaus, Entwicklungsleiter Audi Used Car Platform, AUDI AG

Together with the Comsysto Reply team, as the Product Owner of the Used Car Platform, I was able to develop a modern and flexible sales platform which we brought to go-live within less than 12 months. Thanks to the extensive technical and methodological competence of the Comsysto Reply team, the Used Car Platform is continuously being developed further and will serve us at AUDI AG as the central platform for B2B sales of all commercial and used vehicles in the future. Throughout our long-term collaboration, the team provided a wealth of creative suggestions and efficient solutions – both for technical challenges and process optimizations. Based on my project experience, I would recommend Comsysto Reply to any of my colleagues as a development partner!

(translated from German)

Thomas Kirstein, Product Owner Audi Used Car Platform, AUDI AG


Would you like to find out more about this project or do you still have questions that you would like to clarify? We look forward to talking to you.

Contct us