Casablanca – Morocco is preparing to test the use of artificial intelligence in the monitoring of public procurement, as financial oversight authorities explore whether automated data analysis can help review government contracts, identify potential compliance issues and direct auditors toward files that require closer examination. 

The initiative is being developed by Morocco’s Court of Auditors and regional courts of auditors in partnership with the World Bank. It is currently focused on the development of a prototype rather than the immediate deployment of a nationwide system. Companies interested in developing the proposed technical solution have been invited to submit their proposals, with September 10, 2026, set as the deadline. 

The project reflects a broader effort to use digital tools in public administration and financial oversight as the volume of government procurement data and supporting documentation continues to grow. Rather than replacing existing legal and auditing procedures, the proposed system would initially serve as an analytical tool capable of screening large numbers of files and highlighting cases that may deserve human attention. 

A system designed to examine large volumes of information 

Public procurement procedures generate substantial amounts of information, including contracts, tender specifications, committee minutes, bid evaluations, scoring sheets and other administrative documents. Reviewing these materials manually can require considerable time, particularly when oversight authorities need to examine numerous procurement procedures within a single audit assignment. 

The proposed artificial intelligence system would seek to address this challenge by bringing together structured digital information and unstructured documents. 

Structured information could include dates, contract details, procurement values, deadlines and other data that can be processed automatically. At the same time, the system would be expected to analyze textual documents such as tender specifications and committee reports, extracting information that can be compared with other records. 

The objective would be to identify inconsistencies or indicators that may warrant further investigation. 

For example, an automated system could compare the date a tender was published with the period allowed for submitting bids. It could examine whether requirements contained in tender specifications correspond with the criteria subsequently used to evaluate competing offers. It could also compare information appearing in different documents and flag discrepancies for auditors. 

Such functions could allow oversight teams to conduct an initial review of a much larger number of procurement files than would normally be possible through detailed manual examination alone. 

From sample checks to broader screening 

One of the central ideas behind the project is to broaden the first stage of financial oversight. 

Traditional auditing often requires authorities to select a number of files for detailed examination. The proposed technology could instead allow all contracts falling within the scope of a particular audit assignment to undergo an initial automated screening. 

This would not mean that every contract would receive the same level of detailed scrutiny. Rather, the algorithm could classify files according to the indicators it detects and help auditors determine which cases should receive greater attention. 

Files without significant warning indicators could move through the initial screening process, while those containing unusual patterns, inconsistencies or potentially problematic conditions could be prioritized for further review. 

The approach could therefore change the role of technology within the audit process from simply storing or retrieving information to helping authorities organize and prioritize large amounts of data. 

However, the effectiveness of such a system would depend heavily on the quality and completeness of the information available to it. Missing documents, inaccurate data or inconsistencies in procurement records could affect the reliability of automated analysis. 

Scrutiny of tender requirements 

Tender specifications are expected to be an important area of analysis under the proposed system. 

Algorithms could be trained to identify requirements that potentially restrict competition, unusually specific technical specifications, potentially discriminatory criteria or publication periods that may warrant examination. 

The system could also analyze scoring frameworks and the minutes of committees responsible for opening and evaluating bids. 

However, identifying a potentially restrictive condition would not automatically establish that a procurement procedure had violated the law. A specific technical requirement, for example, may be justified by the nature of a contract or the technical standards required for a particular project. 

For this reason, the proposed technology is designed to generate alerts rather than legal findings. 

The distinction is important because procurement oversight involves assessing the context surrounding individual decisions. An algorithm may identify a pattern that appears unusual, but auditors and financial judges must determine whether there is a legitimate explanation and whether the procedure complies with applicable rules. 

Human oversight remains central 

The proposed model places financial judges and auditors at the center of the decision-making process. 

Artificial intelligence would be responsible for searching, comparing and identifying potential issues, while human officials would verify the underlying information and determine whether further action is justified. 

The system is also expected to explain why a particular alert was generated and identify the data or documents that led to it. This would allow auditors to trace an automated warning back to its original source rather than relying on an unexplained computer-generated conclusion. 

This requirement for explainability is particularly relevant in financial oversight, where findings can have legal and administrative consequences. A warning generated by an algorithm would therefore need to be supported by evidence that can be reviewed independently by the responsible officials. 

The system’s usefulness would consequently depend on more than its ability to process large quantities of information. It would also have to produce sufficiently accurate and relevant alerts to avoid overwhelming audit teams with warnings that do not lead to meaningful findings. 

Prototype stage comes before wider deployment 

The initiative remains at the prototype stage, and any wider implementation would depend on the results of the testing process. 

The September 10 deadline for companies to submit proposals represents the next step in selecting a technical solution capable of meeting the project’s requirements. If a functioning prototype is developed and demonstrates satisfactory results, authorities could assess its potential use in broader oversight assignments. 

A successful system could allow financial oversight bodies to process procurement information more quickly and examine a wider range of contracts during the initial stages of an audit. It could also help auditors identify relationships between information contained in separate documents that might be difficult to detect through manual review alone. 

At the same time, the project raises practical considerations surrounding data quality, system reliability, explainability and the appropriate division of responsibilities between automated tools and public officials. 

For now, Morocco’s approach is centered on testing rather than replacing established oversight mechanisms. The proposed model gives artificial intelligence a supporting role: algorithms would analyze data and documents, identify possible anomalies and issue alerts, while financial judges and auditors would remain responsible for verification, legal assessment and final decisions. 

If the prototype proves effective, the experience could provide Morocco with an additional digital tool for managing the growing volume of information associated with public procurement, while maintaining human review as the basis for financial and legal oversight.