AP’s AI-Led Gst Model Wins Praise at NCM Meet

VIJAYAWADA: Andhra Pradesh’s Commercial Taxes Department stole limelight at the 6th National GST Coordination Meeting (NCM) in New Delhi on Friday, as its Artificial Intelligence (AI) and Machine Learning (ML)-driven tax administration model drew praise.

The presentation, delivered by chief commissioner of state tax, Babu Ahmed, was a dedicated agenda item at the meeting held under the chairpersonship of the revenue secretary to the central government. It was attended by officials from CBIC and GST council secretariat as also commissioners from states and union territories.

The AP taxes department showcased how AI has been embedded across the full spectrum of GST administration — from case selection and allocation to return scrutiny under Section 61, audits under Section 65, inspections under Section 67, and litigation before appellate forums.

A four-source data integration layer, 19 automated analytical reports, and a 35-parameter risk matrix now classify cases before they reach officers, ensuring high-value matters are routed to senior scrutiny committees.

A highlight was the Legal-AI Officer Assistant, trained on the GST statutory corpus and 22,000 judicial precedents. This system has absorbed over 13,700 cases, assisting officers in drafting instructions, para-wise remarks and replies across the First Appellate Authority, GSTAT and high court. Every draft is reviewed and signed by officers, maintaining statutory accountability.

The tax department presented striking comparisons between AI-assisted and legacy processes. It was noted that, in return scrutiny, six months of AI operation yielded revenue detection of Rs 743.43 crore, nearly double the Rs 365.75 crore achieved under manual methods.

Detection per case rose sharply to Rs 27.63 lakh against the Rs 3.08 lakh earlier —a scale of efficiency that was “impossible through manual expansion alone,” officials said.

The tax department stressed that automation remains firmly within statutory workflows: “AI drafts, but officers decide,” with every order digitally signed and traceable through BO Case ID, ARN and DIN.

The presentation drew overwhelming interest from other states. Delegations from Tamil Nadu, Bihar, Rajasthan, Chhattisgarh, Kerala, Assam and Meghalaya sought details of AP’s AI architecture and workflow design. Some of them would plan study visits to Andhra Pradesh.

The central revenue secretary commended the initiative as a model for AI-led tax administration that other states could emulate.

Looking forward, AP outlined plans to extend AI-based risk scoring to provisional refunds, registration monitoring and cancellation, and expand the Legal-AI Assistant to revision petitions, arrears recovery and advance rulings.

The department also urged the Centre and GSTN to enable direct API integration, cross-border taxpayer visibility and sustained peak-period access to fully scale the initiative.

Chief commissioner Babu Ahmed thanked the GST Council secretariat and the revenue secretary for the platform, reiterating AP’s commitment to share its AI framework with other states to strengthen GST administration nationwide.