On February 10, the National Development and Reform Commission (NDRC) issued the "Opinions of the National Development and Reform Commission and other departments on Accelerating the Promotion and Application of Artificial Intelligence in the Field of Tendering and Bidding", which requires that the whole process of tendering and bidding transactions and key links of management should be focused on in accordance with the principles of government guidance, multi-party participation, scenario traction, safety and controllability. We will actively and steadily promote the application of artificial intelligence in the field of tendering and bidding, improve the paradigm of tendering and bidding, enhance the level of digital intelligence in service and supervision, and provide strong support for ensuring the fair and efficient allocation of public resources and standardizing the order of the tendering and bidding market. By the end
of 2026, key scenarios such as tender document detection, intelligent assistant bid evaluation, and identification of surrounding bids will be fully covered in some provinces and cities; By the end of 2027, more key scenarios will be popularized and applied nationwide, and a number of experiences and practices in model training, scenario application and mechanism guarantee will be formed to effectively promote the healthy development of the bidding market.
The full text is as follows:
Implementation Opinions
of the National Development and Reform Commission and Other Departments on Accelerating the Promotion and Application of Artificial Intelligence in the Field
of Tendering and Bidding Development and Reform Regulations [2026] No.
195 All provinces, autonomous regions, municipalities directly under the Central Government, The Sinkiang Production and Construction Corps and Development and Reform Commission, the competent departments of industry and information technology, the Housing and Urban-Rural Construction Department (Commission, Bureau), the Transportation Department (Bureau, Committee), the Water Conservancy (Water Affairs) Department (Bureau), the Agricultural and Rural Department (Bureau, Committee), the competent commercial department, the State-owned Assets Commission, the guidance and coordination department of tendering and bidding, and the integration of public resources trading platform. The communications administrations of all provinces, autonomous regions and municipalities directly under the Central Government, and all central enterprises:
In order to implement the deployment requirements of the CPC Central Committee and the State Council on deepening the reform of tendering and bidding, and in accordance with the relevant arrangements of the Opinions of the State Council on Deeply Implementing the "Artificial Intelligence +" Action, To promote the deep integration of bidding and artificial intelligence and promote the healthy development of the bidding market, the following suggestions are put forward.
First, the overall goal
is to actively and steadily promote the application of artificial intelligence in the field of tendering and bidding, improve the paradigm of tendering and bidding, enhance the level of digital intelligence of service and supervision, and promote the development of the whole process and key links of management of tendering and bidding transactions in accordance with the principles of government guidance, multi-party participation, scenario traction and safety and controllability. The fair and efficient allocation of public resources and the standardization of the order of the bidding market provide strong support. By the end
of 2026, key scenarios such as tender document detection, intelligent assistant bid evaluation, and identification of surrounding bids will be fully covered in some provinces and cities; By the end of 2027, more key scenarios will be popularized and applied nationwide, and a number of experiences and practices in model training, scenario application and mechanism guarantee will be formed to effectively promote the healthy development of the bidding market.
2. Accelerating the application
of scenarios (1) "Artificial Intelligence +" bidding
1. Assisting bidders to conduct comprehensive research and judgment on industry trends, market supply and demand, resource elements, etc., accurately understand and analyze the information of investment approval, bidding, acceptance and acceptance of performance of relevant projects, so as to ensure the quality of projects. Objective and quantitative bidding requirements, technical and commercial conditions are generated to improve the scientific rationality of bidding from the source.
2. On the basis of in-depth understanding of project objectives, construction contents, bidding requirements, technical and commercial conditions, and in combination with historical transactions and policy and regulatory requirements, intelligently match the model bidding documents, recommend appropriate qualifications, bid evaluation methods and standards, and assist bidders in compiling or automatically generating bidding documents. Improve the quality of bidding documents.
3. Combining with the requirements of relevant policies and regulations, the bidding documents are tested in multiple dimensions, such as compliance, rationality and wrong words, to automatically identify all kinds of violations of laws and regulations, exclusion and restriction of competition, to prompt the basis for judgment and suggestions for modification, and to assist the tenderee to correct the bidding documents intelligently. Encourage the implementation of tender documents "first physical examination, then release".
(2) "Artificial Intelligence + Bidding
" 4. Capture all kinds of project bidding information in an all-round way, push suitable project information according to the characteristics of bidders, automatically extract and analyze the key elements of project bidding plan, bidding announcement and bidding documents, intelligently generate the bidding demand map, and provide the bidding information for all kinds of projects. Prompt important content. Combining with historical transactions and similar projects, it assists in analyzing and evaluating the economy of participating in project competition and improving the efficiency of bidding.
5. In-depth analysis of project bidding requirements and requirements of bidding documents, combined with the characteristics and advantages of bidders, to assist bidders to determine the technical scheme and quotation range. The bidding documents prepared by bidders are compared with the bidding documents in terms of responsiveness, which automatically prompts the problems of violation of laws and regulations, errors and omissions in the bidding documents, and assists the bidders to modify and improve them. Risk warning should be given to bids that may be lower than cost price, so as to reduce malicious low-price bidding and other situations.
(3) "Artificial intelligence +" bid opening and evaluation
6. Create humanized digital bid openers, schedule the whole process of project bid opening activities, and automatically complete the work of reading out the bid opening discipline, publishing the list of bidders, deciphering the bidding documents, singing the bidding documents and confirming the results. Intelligently judge the abnormal situation of bid opening and prompt it to assist the tenderee to deal with it efficiently in real time.
7. Comprehensively analyze the characteristics of the project and the key points of bid evaluation, according to the professional classification, geographical distribution, avoidance rules and other conditions of the bid evaluation experts, combined with the requirements of remote bid evaluation, automatically generate the expert extraction scheme suitable for the project, improve the matching degree between the extracted experts and the project, and ensure the scientificity and fairness of the expert extraction.
8. Build humanized evaluation reasoning ability, master the knowledge structure system of different professional bid evaluation experts, establish a comprehensive evaluation index system according to the type of project, combine with specific project recommendation or matching evaluation points, comprehensively extract the elements of bidding documents, deeply analyze the content of bidding documents and the responsiveness of bidding documents. Assist experts to carry out evaluation or generate results for experts' reference, so as to improve the fairness of bid evaluation.
(4) "Artificial intelligence +" calibration
9. Build the ability of intelligent audit of bid evaluation report, assist bidders to verify the data accuracy, logical relevance, content compliance of bid evaluation report, automatically warn the problems of inconsistent scoring of objective evaluation factors, incorrect calculation of scores, excessive deviation of expert scoring, and so on. Prompt the bid evaluation expert to review, confirm, modify and improve.
10. Based on the bidding demand, supply chain management, historical transactions, etc., combined with the data of relevant industry, credit, taxation, judicial and other platform systems, the winning candidates are depicted in a multi-dimensional and three-dimensional way. By introducing digital person defense and other methods, the tenderee is assisted to make a comprehensive comparison and analysis of the winning candidates and make a decision on the bid, so as to record and trace the whole process of the bid.
11. Automatically extract the main signing elements of the winning contract from the bidding documents, refer to the relevant model text and combine with the project-specific contract terms to generate the contract, so as to realize the online signing and archiving of the contract. Combining with the requirements of policies and regulations and historical transactions, the key rights and obligations clauses of the winning contract are prompted to reduce the problems of "Yin-Yang contract" and arbitrary tampering.
(5) "Artificial intelligence +" on-site management
12. Carry out all-round intelligent management of trading venues, efficiently allocate resources and personnel, dynamically monitor all kinds of personnel and activities in trading venues, improve the intelligent level of trading facilities, and create an unattended intelligent trading environment. Strengthen the coordination and complementarity of resources between trading venues, and enhance the facilitation level of cross-regional trading services.
13. Construct a closed-loop witness system of "intelligent research and judgment-dynamic intervention-certificate on the chain", carry out insensitive digital witness for all links of bidding and bidding transactions, and record the whole process of transactions comprehensively and accurately. Strengthen the analysis and early warning of abnormal behavior, promptly remind and dissuade suspected illegal activities, strengthen the report of problem clues, and provide support for the relevant departments to enforce discipline and law.
14. Construct the intelligent management system of bidding and tendering transaction archives, and realize the intelligent filling of transaction documents and the chain storage of transaction data in combination with intelligent witness management. It intelligently names and classifies the bidding transaction data, automatically generates file indexes and abstracts, and provides intelligent retrieval and query services. Fully tap the role of transaction archives in policy performance evaluation, bidding analysis, dispute resolution, cost reduction and efficiency enhancement, so as to enhance the comprehensive utilization efficiency of transaction archives.
15. Build a professional question-and-answer engine in the field of tendering and bidding, provide multi-modal interactive consultation services for various policies and regulations, business knowledge, operation processes, etc., realize the functions of intelligent operation guidance, intelligent recommendation of models, abnormal early warning question-and-answer, objection complaint consultation, etc., and improve the convenience of tendering and bidding transaction services.
(6) "Artificial intelligence +" supervision
16.Construct an intelligent management system for the whole life cycle of bid evaluation experts, combine professional competence, performance assessment, credit evaluation, education and training, carry out multi-dimensional three-dimensional portraits of bid evaluation experts and assist in dynamic assessment, so as to enhance the comprehensive management ability of bid evaluation experts. Combining with the actual situation of the specific industry, it enables the national bid evaluation experts to "one network management" and promotes the sharing and sharing of high-quality expert resources.
17. Construct a comprehensive early warning system with full coverage of "subject + behavior". Through multi-dimensional data collision and subject portraits, we can find hidden problems such as identical enterprise characteristic information, abnormal subject relationship, bidding behavior, bid-winning probability, and expert scoring tendency. The bidding documents, bill of quantities and quotation list are scanned in depth, and the clues of suspected bid-rigging problems are excavated through semantic similarity analysis of technical schemes and comparison of key quotation characteristics of commercial bids, so as to provide reference for relevant departments to enforce discipline and law.
18. Build the ability of intelligent credit management in bidding and tendering, and realize the objective record, automatic collection, shared application and dynamic adjustment of credit information. Intelligent credit evaluation model, multi-dimensional three-dimensional outline of the main credit portrait, accurate and efficient credit evaluation, information push and early warning reminder.
19. Build an analysis and early warning model that runs through the pre-bid, in-bid and post-bid stages of the project, strengthen the collection, management and application of data in the whole process, and through data collision and comparative analysis, automatically identify problems such as non-recruitment, illegal subcontracting, illegal changes in personnel, serious lag in progress, and low, medium and high knots. Strengthen the connection of "execution discipline" in bidding and tendering, realize the intelligent closed-loop management of early warning and transfer of problem clues, coordinated investigation and punishment, and feedback of results, enhance the ability of in-depth analysis and handling of complex cases, and promote the formation of an intelligent supervision pattern of "one network co-governance" of administrative law enforcement, criminal justice, discipline inspection and supervision.
20. Build intelligent complaint handling capability for tendering and bidding, assist administrative supervision departments to analyze complaints, combine policies and regulations, historical cases and investigation and evidence collection, form preliminary review opinions, classify and give handling suggestions, assist in generating complaint handling decisions, and improve complaint handling efficiency. Intelligent screening and handling of malicious complaints will improve the prevention and control of malicious complaints.
3. Standardized deployment and implementation
(7) Scientific organization and promotion. All localities should scientifically determine the implementation path according to the practical needs and technical basis. For scenarios that improve transaction efficiency and are suitable for market-oriented promotion, we should actively cultivate AI application service providers. For the scenario of ensuring fair and just transactions and improving the quality and effectiveness of supervision, we should pay attention to the leading role of the government, strengthen overall planning and intensive construction, deploy and apply prefectures and cities under the overall guidance of the provincial level, and reuse the model resources of superiors at and below the county level in principle.
(8) Strengthen system integration. Local governments should continue to deepen the integration and sharing of public resources trading platforms, carry out intensive transformation in an orderly manner on the basis of unified system rules and technical standards, improve the standardization level of bidding and bidding transaction processes and related platform systems, strengthen the interconnection and interoperability of related platform systems, and improve the efficiency of model deployment and application.
(9) Consolidating the data base. All localities should strengthen the management of bidding data, strengthen data cleaning and labeling, speed up the construction of high-quality data sets and knowledge bases covering all aspects of bidding policies and regulations and the whole process, rely on the government data sharing mechanism, promote the co-construction and sharing of high-quality data sets and the collection and management of generated data, and better support model training and application.
(10) Continuous iterative optimization. All localities should establish a normal upgrade mechanism for artificial intelligence models, update data sets and knowledge bases in a timely manner, use professional bidding data for targeted training, constantly optimize the model algorithm, and improve the accuracy of the model. It is necessary to establish a feedback mechanism for user evaluation, collect and process user needs in time, improve application functions, and drive iterative optimization of the model with user feedback.
(11) Improving the application mechanism. All localities should strengthen the connection between the application of artificial intelligence and the whole process of bidding and tendering, improve the transformation and application mechanism of model generation content, and ensure that the model plays its full role. Adhering to the auxiliary orientation of technology, the conclusion generated by the model does not replace the independent judgment of bidders, bidding agencies, bidders, bid evaluation experts, etc., and does not change the legal responsibility of the user.
(12) Improve the level of safety. Strictly implement the security management requirements of artificial intelligence models, strengthen the security capacity building of model algorithms, data resources, infrastructure and application systems, and strictly carry out algorithms, model filing and security audits. Construct data, computing power, algorithm and system security protection system to ensure the safety and reliability of the model, and effectively prevent and deal with the risks of model black box, illusion and algorithm discrimination.
4. To strengthen organizational guarantee
, provincial development and reform departments should effectively play a guiding, coordinating and leading role, intensify organizational implementation efforts, actively coordinate and solve data and computing needs, jointly with relevant departments to determine application scenarios and implementation paths as soon as possible, promote implementation by classification, and improve the application guarantee mechanism; We should strengthen cooperation with universities, research institutes and artificial intelligence enterprises, give full play to the role of artificial intelligence enterprises, promote the transformation of production, teaching and research, strengthen the construction of talent team, and strengthen the training of cross-disciplinary talents. The National Development and Reform Commission will strengthen overall coordination with relevant departments, do a good job in publicity, guidance and risk control, guide local and central enterprises to deepen exploration and application according to local conditions, improve supporting systems and rules, promote the construction of standard system, and timely summarize and popularize typical experiences and practices.
浙公网安备33010802003254号