With new businesses multiplying in the Shenfu reform demonstration zone, tax risks have grown more varied, so the local tax bureau straightened out its people, data and processes at once. About 30 staff were split into a task-coordination group that merges and de-duplicates leads, a data group that analyses figures and validates models, and a field group that verifies on site; findings and newly discovered indicators flow back to improve the models. The core tool is a self-built risk-judgement model that cleans high-frequency indicators from filings, payments and invoices, tags them by industry, scenario and company type, and pushes matching indicators automatically, so risk work follows rules instead of intuition. One high-tech materials company was flagged for abnormal inventory revenue and unpaid stamp duty on transport invoices, and verification showed it was claiming high-tech incentives despite falling short on R&D spending; it paid the tax plus late fees.
News Detail
Shenfu demonstration zone tax bureau builds a three-pronged risk defense
The Shenfu reform demonstration zone tax bureau split its risk work across three teams for task coordination, data analytics and field verification, backed by a self-built risk-judgement model. The model cleans high-frequency indicators from filings, payments and invoices, tags them by scenario, company type and industry, and pushes the right indicator sets automatically, shifting risk work from experience-based searching to rule-based matching. Findings flow back to refine the model, and one case caught a high-tech enterprise wrongly claiming preferential treatment.
2026-10-26