Industry Ministry Smart Project Reaches 70 Percent Completion Amid Data Access Hurdles
According to the Economic Desk of Webangah News Agency, the Isfahan University of Technology is spearheading the development of the “Intelligent Assistant for the Ministry of Industry, Mine, and Trade” (MIMT), following a mandate from the Vice Presidency for Science and Technology. Alireza Basiri, the project manager, confirmed that the artificial intelligence platform has already been deployed on a trial basis within the Ministry’s legal and regulatory framework and is currently in use by the Ministry’s legal office.
The MIMT AI platform is engineered around core functionalities including legal analysis, business intelligence, management dashboards, and data interaction through natural language processing, aiming to operationalize AI in the Ministry’s management and decision-making levels.
Despite the 70 percent completion mark, Basiri stressed that the most significant obstacle to achieving the system’s complete maturity remains the critical challenge of accessing unified and actualized data. He noted that the data currently provided by the MIMT is limited, necessitating work with larger, more complex real-world data sets for the system to reach its full potential.
Basiri detailed that the “Intelligent Assistant” project commenced seriously in Esfand 1403 (March 2024) after the proposal, developed by a team including six faculty members from Isfahan University of Technology’s Electrical and Computer Engineering department, was approved by the Science and Technology Vice Presidency. Approximately twenty personnel, comprising academic staff and the executive team, are directly engaged in the development.
The project roadmap, defined by the Vice Presidency, is structured in three to four distinct phases. The initial phase, which has been fully implemented and validated by the Ministry’s supervisor, focused on refining and retrieving MIMT regulations. This phase successfully integrated capabilities for processing and analyzing legal texts, allowing users to query legal content and identify conflicts or contradictions between specific articles and existing regulations. The legal section is now actively utilized by the Director-General of Legal Affairs at MIMT.
Phase two involves the development of the Ministry’s Business Intelligence (BI) component, which is currently about 70 percent complete, covering both static and dynamic BI modules. Basiri explained that classical BI consolidates disparate data systems within an organization into a unified reference source for senior management, moving beyond the operational focus of individual information systems like the Cadastre or Pargar systems. The goal is to generate aggregated reports, facilitated by integrating various databases into a central data warehouse with defined validation processes, which feed into management dashboards.
In terms of dynamic BI, managers can interact directly with the system using natural language queries—for instance, asking for the number of mines categorized by province—without needing to write code or use traditional reporting tools. The system automatically identifies necessary tables, executes the required SQL queries, and presents the results in both tabular and graphical formats. These dynamic queries can then be saved as static, recurring reports on the BI dashboard.
Basiri reiterated that the primary bottleneck is data access, emphasizing that the complexity of integrating and utilizing larger, real-world data sets presents new challenges compared to initial testing with simpler data. The project has passed the Minimum Viable Product (MVP) stage and is entering operational testing.
Beyond the legal and BI modules, other features requested by the Science and Technology Vice Presidency, or deemed crucial for practical use, have been integrated. These include secure text summarization capabilities, ensuring confidential Ministry documents are processed internally rather than relying on external public AI chats. Translation functionality for nearly 100 languages has also been added, alongside the ability for users to upload and query non-integrated data sources, such as large CSV files, using the same analytical tools.
The total allocated budget for the project entrusted to Isfahan University of Technology is 15 billion Toman. The university leveraged its established computational infrastructure, including a supercomputer established in collaboration with Amirkabir University of Technology in 1389 (2010), to manage the heavy GPU demands. Although the project did not explicitly cover hardware procurement, the university provided necessary resources, including access to its existing GPU clusters since Farvardin (March/April). The system is designed with fault tolerance, automatically switching from the primary processing platform—the National AI Platform hosted by Sharif University since Mordad/Shahrivar (July/August)—to the Isfahan University platform if connectivity fails.
