- Andhra Pradesh's AI and ML-driven tax administration model won national acclaim at the 6th National GST Coordination Meeting (NCM) in New Delhi.
- The model was presented by Chief Commissioner Babu Ahmed as a dedicated agenda item at the meeting.
- The central revenue secretary praised the initiative as a model for other states to emulate.
- Delegations from Tamil Nadu, Bihar, Rajasthan, Chhattisgarh, Kerala, Assam, and Meghalaya sought details of AP's AI architecture and workflow design.
- AP plans to extend AI-based risk scoring and expand the Legal-AI Assistant to enhance its GST administration.
- The AI and ML-based system developed by AP covers the entire GST administration process, including case selection, return scrutiny, audits, inspections, and litigation.
- The system integrates multiple data sources, automated analytical reports, and a risk matrix to identify cases requiring scrutiny.
- In just six months, the AI-driven return scrutiny led to a revenue detection of ₹743.43 crore, nearly double the amount detected under manual methods.
- The Legal-AI Officer Assistant has been trained on GST laws and around 22,000 judicial judgments to assist officers in litigation.
- The department emphasized that automation remains within statutory workflows, ensuring that AI assists but does not replace officer decision-making.
Andhra Pradesh's innovative AI-powered GST administration model garnered national acclaim at the 6th National GST Coordination Meeting, showcasing its effectiveness in enhancing tax collection and efficiency.1
Chief Commissioner of State Tax Babu Ahmed presented the model, which integrates a four-source data layer, 19 automated reports, and a 35-parameter risk matrix to streamline case selection and scrutiny.
The system has transformed the tax administration process, covering everything from audits to litigation. A standout feature is the Legal-AI Officer Assistant, trained on GST laws and 22,000 judicial precedents, which has processed over 13,700 cases.9
Six months of AI-driven return scrutiny resulted in a remarkable revenue detection of ₹743.43 crore, nearly doubling the ₹365.75 crore detected through manual methods. Detection per case surged to ₹27.63 lakh, compared to ₹3.08 lakh previously, demonstrating a significant leap in efficiency.8
Delegations from states including Tamil Nadu, Bihar, and Rajasthan expressed interest in replicating Andhra Pradesh's model, which emphasizes that while AI assists, final decisions remain with human officers.
The central revenue secretary's endorsement of this initiative as a model for other states underscores its potential impact on tax administration nationwide.3
“The AI system, trained on GST laws and 22,000 judicial precedents, has absorbed over 13,700 cases, assisting officers in drafting replies across appellate forums. In six months, AI-driven return scrutiny detected ₹743.43 crore in revenue, nearly double the ₹365.75 crore under manual methods.”
Deccan Chronicle
