On June 22, 2026, the Inner Mongolia tax authority said it had built a foundation combining tax-and-fee relationship graphs with a business-scenario knowledge base on top of its tax data platform. It mapped 110 correlation rules across 37 tax and fee categories and embedded large-model capabilities into risk analysis, indicator generation, script development, and quality review. Public information shows the application has expanded from single-point anomalies such as mismatches between invoice items and business scope to VAT filing checks and full-chain industrial risk control. In development, the model can turn an idea into a script within minutes, and in review it can complete multi-dimensional checks and produce an assessment report with issue-recognition rates above 99 percent. The case shows tax informatization moving toward automated rule generation, earlier risk detection, and higher-quality verification.
News Detail
Inner Mongolia Tax Uses Large Models to Improve Tax Risk Management and Review Quality
Inner Mongolia tax authorities are using large models to improve risk management and review quality by linking rule mapping, script generation, quality checks, and frontline verification into one intelligent risk-control workflow.
2026-08-19