Request a Demo
Case studies
Red Flags Management: Aiding the General Prosecutor's Office in Detecting Risky Public Procurements
The General Prosecutor's Office exercises the highest supervision over the observance of legality in Kazakhstan, including public procurement supervision. Every year, the state concludes over 1.5 million contracts, and controlling compliance with all the requirements is impossible to perform manually. At the same time, timely identifying transactions with signs of risks and violations is very important, as we are talking about $25 billion a year that the state receives from taxpayers and must spend wisely on society's needs. Unfortunately, there are always those who want to take advantage of their role in the procurement process and use part of this amount for their selfish purposes.
In order to help the General Prosecutor's Office identify and timely prevent financial risks, Datanomix has implemented Red Flags Management, an analytical solution that automatically finds risky tenders and contracts, focusing the auditors’ attention on objects with the highest risk of inefficient use of funds.
In order to help the General Prosecutor's Office identify and timely prevent financial risks, Datanomix has implemented Red Flags Management, an analytical solution that automatically finds risky tenders and contracts, focusing the auditors’ attention on objects with the highest risk of inefficient use of funds.
ABOUT SOLUTION
Procurement Red Flags Management solution works by leveraging data analytics, business intelligence, and artificial intelligence (AI) to streamline the identification and evaluation of risky procurements.
The system capability includes:
Advanced algorithms and AI techniques help to automatically identify red flags in procurement processes. These red flags could indicate potential fraud, corruption, or inefficiencies in the process. Examples of red flags include overpricing, single-source contracts, unusual bid patterns, conflicts of interest, unusually high or low bids, change orders or contract amendments.
The system calculates risk scores for each procurement based on the identified red flags, taking into account factors such as the value of the contract, the complexity of the procurement, and the potential impact of the risk. Focusing the attention of auditors on the riskiest procurements enables auditors to prioritize their efforts on high-risk procurements.
The solution provides user-friendly dashboards, reports, and visualizations to help auditors and procurement professionals easily review and analyze the identified risks, prioritize actions, and track progress.
Procurement Red Flags Management solution works by leveraging data analytics, business intelligence, and artificial intelligence (AI) to streamline the identification and evaluation of risky procurements.
The system capability includes:
- Procurement control
- Cost analysis for the fungible commodities
- Performance analysis of the procurement process
- Competitiveness analysis by different categories
- Analysis of the procurement processes duration by categories
- Procurement benchmarking
Advanced algorithms and AI techniques help to automatically identify red flags in procurement processes. These red flags could indicate potential fraud, corruption, or inefficiencies in the process. Examples of red flags include overpricing, single-source contracts, unusual bid patterns, conflicts of interest, unusually high or low bids, change orders or contract amendments.
The system calculates risk scores for each procurement based on the identified red flags, taking into account factors such as the value of the contract, the complexity of the procurement, and the potential impact of the risk. Focusing the attention of auditors on the riskiest procurements enables auditors to prioritize their efforts on high-risk procurements.
The solution provides user-friendly dashboards, reports, and visualizations to help auditors and procurement professionals easily review and analyze the identified risks, prioritize actions, and track progress.
«Over 18 thousand procuring entities conclude more than one and a half million of public procurement contracts in Kazakhstan annually. Earlier inspections were carried out selectively based on internal intelligence, according to complaints or appeals from suppliers. It wasn’t very efficient. Now we can use the Red Flags Management solution to analyze all contracts on the public procurement portal and check only the high-risk ones», - reported the press service of the Prosecutor General's Office.
PILOT PROJECT
Before the implementation of the RFM system, prosecutors directly worked with the public procurement portal (goszakup.gov.kz), which publishes all data on the state procurement process in Kazakhstan, from planning to payment to the supplier.
Datanomix has integrated the RFM system with the procurement portal, using the public API. The RFM system receives a large amount of data on Kazakhstan's government procurement - about 100 million records - on a daily basis and provides prosecutors with advanced capabilities to analyze procurement and prevent illegal transactions.
Datanomix's risk detection algorithms are based on international experience (World Bank, OECD, Open Contracting Partnership) and adapted to the requirements of the current legislation of the Kazakhstan public procurement, as well as to national specifics.
Datanomix has integrated the RFM system with the procurement portal, using the public API. The RFM system receives a large amount of data on Kazakhstan's government procurement - about 100 million records - on a daily basis and provides prosecutors with advanced capabilities to analyze procurement and prevent illegal transactions.
Datanomix's risk detection algorithms are based on international experience (World Bank, OECD, Open Contracting Partnership) and adapted to the requirements of the current legislation of the Kazakhstan public procurement, as well as to national specifics.
After integration and configuration of risk detection algorithms, we conducted a demonstration of the system for the Deputy General Prosecutor. At first, he was skeptical about the data analysis capabilities in the fight against corruption. But during the demonstration, we and other prosecutors began to look at risky procurements that our system had identified. The cases were interesting, and at the demonstration, it was decided to conduct an inspection for these procurements and pilot the system. Later, the identified transactions turned out to be dishonest.
Vitaliy Trenkenshu
CEO of redflags.ai
During the pilot project, Datanomix consultants and General Prosecutor's Office prosecutors worked together closely. A special chat was created for operational communication, where consultations, training, and brainstorming meetings were held to discuss new risk ideas. Prosecutors shared their ideas and practical experience regarding various public procurement criminal schemes and violations. Thanks to this, Datanomix consultants were able to develop new risk algorithms that help prevent offenses in Kazakhstan.
"The most frequent violations are related to the improper way of conducting purchases, their announcement without design and estimate documentation, when it is mandatory, the presentation of excessive requirements that exclude a competitive environment. Particular attention is paid to identifying facts of inefficient budget planning," representative of the prosecutor's office of the Abay region shares his observations.
In total, the data analysis methodology for government procurement contains 42 risks at all stages of the procurement process, from planning to payment to the supplier. Particular attention is paid to the planning stage, as it is always better to prevent an offense than to detect it after the fact.
INTEGRATIONS
At the next stage, additional data sources were connected to the system, allowing for a more comprehensive analysis of public procurement. In addition to data from the public procurement portal of Kazakhstan, the RFM solution received access to:
At the next stage, additional data sources were connected to the system, allowing for a more comprehensive analysis of public procurement. In addition to data from the public procurement portal of Kazakhstan, the RFM solution received access to:
- Tax data of legal entities. The system uses data on taxes paid by legal entities to check the reliability of suppliers and look for signs of shell companies.
- Database of certificates of goods origin. According to the rules of public procurement in Kazakhstan, for some groups of goods there are preferences for local producers. However, unscrupulous vendors often forge such certificates. Therefore, for each contract, RFM solution automatically checks whether the supplier had a valid certificate at the time the contract was concluded.
- Register of unscrupulous suppliers. Some suppliers are not allowed to enter into public procurement contracts. The system automatically checks if the supplier is on the list of dishonest suppliers.
TRAININGS
Access to the system was given to prosecutors of all administrative-territorial units of Kazakhstan (the prosecutor's offices of 17 regions and 3 cities of republican significance). Datanomix experts conducted the training where prosecutors learned about new analysis methods and algorithms, as well as gained practical skills in working with the RFM solution. In addition, video and text instructions were developed. A dedicated Datanomix consultant was assigned to the project to support, train and develop new risk algorithms. A special messenger channel allowed prosecutors to quickly receive answers to any arising questions related to working with the solution.
Access to the system was given to prosecutors of all administrative-territorial units of Kazakhstan (the prosecutor's offices of 17 regions and 3 cities of republican significance). Datanomix experts conducted the training where prosecutors learned about new analysis methods and algorithms, as well as gained practical skills in working with the RFM solution. In addition, video and text instructions were developed. A dedicated Datanomix consultant was assigned to the project to support, train and develop new risk algorithms. A special messenger channel allowed prosecutors to quickly receive answers to any arising questions related to working with the solution.
RESULTS
Datanomix's RFM solution helps prosecutors in Kazakhstan identify economic and corruption risks in government procurement. It is impossible to separate the results of the analytical system from the results of human actions, since preventing illegal actions is a joint work of man and machine. The system provides prosecutors with hints, drawing their attention to certain risks, but only a human can make a decision on what actions should be taken (if any).
The effectiveness of the solution can be judged by the following figures: in the first year of using RFM (2017), prosecutors identified 37 illegal schemes worth a total of $108 million. In 2021, prosecutors canceled 505 illegal tenders worth a total of $331 million.
Having worked for many years in the field of government procurement analysis, the Datanomix team can state that the Kazakhstan government is actively improving government procurement procedures to increase their efficiency and reduce the number of violations. Therefore, the impressive increase in the number of prevented illegal tenders is undoubtedly associated not with an increase in crime while maintaining the percentage of detected offenses, but due to more effective detection of offenses, including thanks to the technologies used. Thanks to RFM the prosecutors suppress such violations that were previously almost impossible to detect.
Datanomix's RFM solution helps prosecutors in Kazakhstan identify economic and corruption risks in government procurement. It is impossible to separate the results of the analytical system from the results of human actions, since preventing illegal actions is a joint work of man and machine. The system provides prosecutors with hints, drawing their attention to certain risks, but only a human can make a decision on what actions should be taken (if any).
The effectiveness of the solution can be judged by the following figures: in the first year of using RFM (2017), prosecutors identified 37 illegal schemes worth a total of $108 million. In 2021, prosecutors canceled 505 illegal tenders worth a total of $331 million.
Having worked for many years in the field of government procurement analysis, the Datanomix team can state that the Kazakhstan government is actively improving government procurement procedures to increase their efficiency and reduce the number of violations. Therefore, the impressive increase in the number of prevented illegal tenders is undoubtedly associated not with an increase in crime while maintaining the percentage of detected offenses, but due to more effective detection of offenses, including thanks to the technologies used. Thanks to RFM the prosecutors suppress such violations that were previously almost impossible to detect.
"The RFM software has made a significant contribution to reducing risks in government procurement at the stages of planning purchases and contractual relations and has allowed to increase the level of analysis conducted by the General Prosecutor's Office staff," comments Rizat Bolatovich Smagulov, Deputy Head of the Public Interest Protection Service.
KEY LESSONS LEARNED
As our joint work has shown, detecting potential risks and working to prevent them is much more effective than dealing with the consequences. Therefore, the risk detection system at the planning stage is so effective, as the prosecutor's office can intervene and eliminate them before any damage is done to the state.
It is even better to eliminate the causes that contribute to these risks. To eliminate some risks, pinpoint changes are required in the legislation on government procurement, procurement rules, or the electronic government procurement portal. The RFM system allows analyzing statistics on risks, seeing trends and bottlenecks that the General Prosecutor's Office of Kazakhstan can use as a basis for recommendations to the government.
It is also important to note that prosecutors have many other work tasks besides monitoring public procurement. The work effectiveness directly depends on the availability of dedicated employees who are engaged only in analytical work. This year, the General Prosecutor's Office of Kazakhstan plans to allocate a separate group of employees to work with the RFM system. Of course, this will further increase the effectiveness of detecting and suppressing risky transactions.
As our joint work has shown, detecting potential risks and working to prevent them is much more effective than dealing with the consequences. Therefore, the risk detection system at the planning stage is so effective, as the prosecutor's office can intervene and eliminate them before any damage is done to the state.
It is even better to eliminate the causes that contribute to these risks. To eliminate some risks, pinpoint changes are required in the legislation on government procurement, procurement rules, or the electronic government procurement portal. The RFM system allows analyzing statistics on risks, seeing trends and bottlenecks that the General Prosecutor's Office of Kazakhstan can use as a basis for recommendations to the government.
It is also important to note that prosecutors have many other work tasks besides monitoring public procurement. The work effectiveness directly depends on the availability of dedicated employees who are engaged only in analytical work. This year, the General Prosecutor's Office of Kazakhstan plans to allocate a separate group of employees to work with the RFM system. Of course, this will further increase the effectiveness of detecting and suppressing risky transactions.
NEXT STEPS
Currently, the system only uses open data sources and has already helped achieve impressive results. However, its capabilities are not exhausted. Closed data sets, such as information about company founders, invoices, information about family ties, and others, can be used to identify many risks.
In 2023, the plan is to transfer the system from Datanomix servers to the secure network of the General Prosecutor's Office. This will allow more data to be loaded into it from various government databases, as the General Prosecutor's Office has the authority to access this data, and the data will be securely protected on its servers without connecting to the Internet. This will allow for better identification of affiliation between customer employees and suppliers, the development of scoring based on signs of fictitious business activity to identify risky suppliers, and so on.
Another area of development is working with unstructured data. For many risks, it is necessary to analyze textual unstructured information from technical specifications. The main way to restrict competition in government procurement remains the specificity of requirements for a particular supplier. Large Language Models (analogous to GPT-4, trained for a specific task on data from the government procurement portal) will be used to identify such intentional obstacles to fair competition. This will allow analyzing the texts of any procurement documentation however they are formulated.
Currently, the system only uses open data sources and has already helped achieve impressive results. However, its capabilities are not exhausted. Closed data sets, such as information about company founders, invoices, information about family ties, and others, can be used to identify many risks.
In 2023, the plan is to transfer the system from Datanomix servers to the secure network of the General Prosecutor's Office. This will allow more data to be loaded into it from various government databases, as the General Prosecutor's Office has the authority to access this data, and the data will be securely protected on its servers without connecting to the Internet. This will allow for better identification of affiliation between customer employees and suppliers, the development of scoring based on signs of fictitious business activity to identify risky suppliers, and so on.
Another area of development is working with unstructured data. For many risks, it is necessary to analyze textual unstructured information from technical specifications. The main way to restrict competition in government procurement remains the specificity of requirements for a particular supplier. Large Language Models (analogous to GPT-4, trained for a specific task on data from the government procurement portal) will be used to identify such intentional obstacles to fair competition. This will allow analyzing the texts of any procurement documentation however they are formulated.
MORE INFO
The Prosecutor General's Office of Kazakhstan has achieved impressive economic results using the Red Flags Management solution to analyze public procurement and stop illegal transactions. Focusing on high-risk cases and facilitating timely corrective actions, Red Flags Management enables it to address fraud and corruption risks, enhance compliance, and improve procurement efficiency. If you want to learn more about the RFM, please contact us.
The Prosecutor General's Office of Kazakhstan has achieved impressive economic results using the Red Flags Management solution to analyze public procurement and stop illegal transactions. Focusing on high-risk cases and facilitating timely corrective actions, Red Flags Management enables it to address fraud and corruption risks, enhance compliance, and improve procurement efficiency. If you want to learn more about the RFM, please contact us.
Resources
Company
See the advantages of using our platform for your procurement
See our award-winning solutions in Action
Recognized by
