Bizz Buzz Bureau Amaravati

AndhraPradesh's Commercial Taxes Department on Friday showcased its Artificial Intelligence (AI) and Machine Learning (ML) led GST administration model at the 6th National Coordination Meeting (NCM) in Delhi, winning huge appreciation from all..

The meeting was held under the chairpersonship of the Union Revenue Secretary, attended by the GST Council Secretariat, the Central Board of Indirect Taxes and Customs (CBIC), and Commercial Tax Commissioners and senior officers from States and Union Territories across the country.

Andhra Pradesh presented data-backed comparisons of AI-assisted outcomes against the legacy, manual process. In return scrutiny alone, six months of AI-driven operation yielded revenue detection of Rs743.43 crore against Rs365.75 crore under the comparable legacy period, and detection of Rs27.63 lakh per case against Rs3.08 lakh per case under the legacy process — an outcome the department noted could not have been achieved through manual scaling alone.

The Department emphasised that automation has been built strictly within the statutory workflow: AI drafts, but officers decide, with every notice and order carrying an officer’s digital signature and full traceability through BO Case ID, ARN and DIN generation.

Andhra Pradesh's presentation, titled “Use of AI/ML and Automation in GST Tax Administration —From Case Selection to Litigation,” was taken up as a dedicated agenda item on the use of AI to improve compliance, taxpayer services and tax administration, and was delivered by A Babu, Chief Commissioner of State Tax, the afternoon session of the meeting.

The presentation set out how the department has built AI-based systems across the full lifecycle of tax administration — case selection and allocation, return scrutiny under Section 61, GST audit under Section 65, GST inspection under Section 67, and litigation before the First Appellate Authority, the GST Appellate Tribunal (GSTAT) and the High Court.

Officers detailed how a four-source data integration layer, 19 automated analytical reports and a 35-parameter risk matrix now screen and classify cases before they reach an officer, with statutory value-based routing to Joint Commissioners, RIMC and SIMC ensuring high-value cases receive additional scrutiny.

The department also demonstrated its Legal-AI Officer Assistant, an AI system trained on the complete GST statutory corpus and approximately 22,000 judgments from the Supreme Court, High Courts, Tribunals and the First Appellate Authority.

The same engine has been extended, without separate development, across three litigation forums — the First Appellate Authority, GSTAT and the High Court (through the Online Legal Case Management System) — absorbing over 13,700 cases into AI-assisted drafting of instructions, parawise remarks and replies, with every draft reviewed and signed off by an officer before issuance.