Machine learning makes research processes more efficient
Using a new process supported by machine learning, the editorial team of the DTAD platform can research and provide relevant order information from the public and commercial sectors for its customers in a more efficient way. Curator is a new internal product that DTAD and codecentric developed together.
DTAD GmbH operates the DTAD platform, a digital tool that helps more than 5,500 active customers a day find new assignments in the public and commercial sectors. The DTAD platform makes it easier for companies to access public tenders and qualified B2B leads, and also helps them to manage these effectively during the acquisition process.
The project at a glance
- Task-based working methods provide focus
- An AI-supported process takes the pressure off the editorial team
- Satisfied users who enjoy working with Curator
- A flexible new cloud system reduces costs
Initial situation
DTAD GmbH's 
IT system has grown steadily over the last 20 years. The editorial team's tasks include researching details of tenders, construction projects, buildings, and contact persons for planned and published projects. During this time, the way the editorial team works and the needs of our customers have changed. The system has adapted accordingly, and technical legacy issues have arisen in some cases.
The internal project Curator MVP (minimum viable product) was established to quickly create a new and sound foundation. It needed to meet current and future customer requirements at the technical and process levels. Specifically, this means that, for example, a construction project used to involve a process consisting of various tools and people. Editorial staff were assigned to the tasks manually.
One thing that quickly became clear in the project was that a new solution could not be developed from scratch, because day-to-day business needed to continue without interruption. The 30 editorial team members had to be able to continue working with as little additional effort as possible, and the more than 5,500 customers had to be able to continue viewing new tenders and commercial contract opportunities. Realizing this project alone is very challenging, which is why DTAD GmbH chose codecentric as its partner.
Solution
In the first step, we conducted a design sprint together with our partner UX&I. We then worked with stakeholders and users to develop the scope for the minimum viable product (MVP). The purpose of the MVP was to provide the editorial team with a web-based and prioritized to-do list for construction projects with contact persons who still had to be researched. The entire process is task-based. This means that research tasks are created automatically; machine Learning supports the editorial staff at various points. To this end, data was analyzed and a model was trained to prioritize tasks in order to research construction projects that are of interest to customers.
Development was carried out by a mixed team consisting of DTAD and codecentric employees. Here's how it was implemented in technological terms: Construction projects are published from the on-premises inventory system to a newly created Amazon Web Services (AWS) environment with an event-driven architecture (EDA) in order to be processed further. AWS Lambda, SNS (Simple Notification Service), and SQS (Simple Queue Service) are used for processing in the cloud. These concepts and services offer the advantage of flexible, scalable, and cost-efficient event processing.
Result
After five months of development, the editorial team was able to incorporate Curator into its daily operations. New construction projects with research by project participants are now processed in the new system and can be curated more easily by the editorial staff. This acceleration of workflows benefits DTA's customers, who are presented with interesting construction projects even faster.
The project enabled us to successfully bridge the gap between the existing and new systems. The editorial team was able to continue working without interruption during the conversion with little additional effort. Users of the DTAD platform continued to be kept informed about construction projects without a hitch. The system is also designed to be so flexible that DTAD can implement changes quickly and easily.
The Curator team cooperated excellently, enabling them to achieve the goal together, in an iterative and user-oriented manner. In addition, focus was placed on the truly essential “M” of MVP for the launch. The joint approach to machine learning has been successfully integrated into the organization's processes, and an ML model is already showing that construction projects can be accurately predicted based on three customer views. Other areas also show that the approach has great potential for further optimization. We will realize this potential with DTAD in follow-up projects.
Kai Noffke
Head of Technology, DTAD GmbH
Our codecentric colleagues have outstanding expertise, are highly competent, very experienced, and incredibly fast. It was really fun to work together. The results are also impressive, since our editors can now process the right construction projects even faster and easier, thereby increasing the value provided for our customers.
Kai Noffke
Head of Technology, DTAD GmbH
Any questions about the project?
Are you interested in a custom cloud-based solution for your business? Let's meet and talk.
People Lead & Agile Consultant
Further projects of codecentric AG
Find out about further successful projects that we have completed with our customers. Perhaps you will find inspiration for a use case in your company.
Sebastian Steiner
People Lead & Agile Consultant