The project
Automated system for assessing the reliability of contractors
Year
2024-2025
Client
One of the largest developers in Russia with a construction volume of more than 1 million m².
The situation at the start of the project
More than 10,000 contractors are in two large registers: construction and application contractors. The registers include various industries: from landscaping and installing playgrounds to cleaning.
An attempt to automate the verification process has already been made: the Client's internal team started developing the system, but the deadlines were constantly shifted due to conflicts between resources and other projects.
Issues
The company works with a large number of counterparties, each of which is manually verified. This is especially critical in the construction industry: the high risk of missed deadlines and contractor bankruptcies can lead to a recurring scenario: an advance is paid, the contractor has defaulted on its obligations, deadlines are missed, and the process is started anew. This entails both direct financial losses and indirect losses — in the form of delayed commissioning of facilities.
Each test required manual analysis of multiple disparate data from different systems. This approach increased the workload on employees, increased the likelihood of mistakes and slowed down the decision-making process.
Business result
- The contractor evaluation methodology, which required manual review, has now been automated. Intermediate stages take place without human intervention, and critical decisions take place in coordination with an expert.
- Problem contractors are identified at an early stage, and this directly affects construction efficiency and reduces the risk of missed deadlines.
- The process has become not only more accurate, but also more convenient: data from different systems is now available in one window, without the need to manually collect information.
- Previously, an assessment required a team of 15 people, but now three people are up to the task — through automation and a single digital environment.
{{cta}}
What is special about the project?
“The peculiarity of this project is that we got into the process after the start: the internal team started development, but then the priorities changed and we took the job ourselves. We worked out the requirements along the way — together with the customer, diving into the business context and understanding the details. As a result, we brought the system to full operation. The project turned out to be lively, dynamic and truly collaborative.”
Project manager on the part of KT.team
Development process
Business request at the start
Automate the contractor evaluation methodology developed by the internal team.
The goal of the project at the start
Create a system that will ensure that contractors are tested at various stages of interaction with the Client:
- When registering a new counterparty.
- When applying for participation in a tender.
- When requesting uniform prices for construction materials.
- When requesting the purchase of materials.
Development time
August 2023 — project launch on the client side.
August 2024 — KT.team joins the project.
May 2025 — the system implemented 8 full integrations with the Client's internal systems.
Result
A system has been developed in which all applications for evaluating contractors are aggregated and displayed in the form of a list with the ability to view:
- risk factors found;
- reliability indicators;
- key information on the basis of which the decision is made.
For the initial assessment, it is enough to open the application card and agree (or disagree) with the automatically generated decision. All interim checks take place without human intervention: attention is required only when suspicious signs are detected.
The system aggregates the following data on the contractor:
- Data obtained by TIN from Kontur.Focus.
- Internal data on open court cases, concluded contracts and receivables.
- Getting into blacklists, which are manually maintained and regularly updated.
The system is integrated with five internal data sources and receives applications from various business processes of the Client.
Technology stack
- Python with the Django framework is a backend with a convenient admin panel.
- React is a modern, fast frontend.
- Kafka — for reliable integration with internal systems and processing application flows.
{{cta}}
