It is hard to conduct data docking. Cross‑department data interfaces are inconsistent, unstructured data takes up a large proportion, the cost of data cleaning is high and the traceability of data quality is tough. Data from all regions shall be consistent with the data delivered to the central and provincial authorities. In response to actual business demands, fields for data governance also need to be expanded in practice.

Data identification is difficult. Massive unstructured data is generated from businesses including supervision coordination and supervision implementation, which creates obstacles to data governance.

Data collection is difficult. Case‑handling data features high‑level permission control and strong isolation, and traditional data‑management methods are inapplicable.

First, corrupt behaviors mutate rapidly. Supervision measures and practical experience for anti‑corruption work require constant iteration.

Second, multiple investigation tactics are adopted. Party‑building and integrity‑oriented education needs to reach front‑line staff and relevant experience awaits upgrading.

Third, staff members of disciplinary inspection departments are under heavy pressure. Repetitive manual work needs to be eliminated and professional experience needs to be updated.

Fourth, the law‑enforcement procedure calls for supervision so as to safeguard the basic rights of involved personnel and achieve experience‑based compliance.

First, computing‑power demands in the AI era fail to match construction conditions. Considering the specialty of disciplinary inspection work, data processing must be completed over dedicated networks and local‑area networks, where real‑time allocation of limited computing resources cannot be realized.

Second, performance gaps exist between digital‑intelligent systems and domestic‑made substitute equipment. Adaptive iteration shall be carried out throughout the whole construction phase.