The tax department is increasingly using artificial intelligence (AI), data analytics and information from multiple government databases to identify anomalies, detect potential tax risks and sharpen scrutiny of taxpayers.
According to tax officials and chartered accountants TNIE spoke to, the approach is moving beyond checking individual tax returns to analysing patterns across Income Tax Returns, GST returns, e-invoices, e-way bills, registrations and historical taxpayer behaviour. The objective is to make scrutiny more targeted and improve voluntary compliance.
The Income-Tax department already processes information available through systems such as the Annual Information Statement (AIS), which gives taxpayers details of financial transactions and income-related information available with the tax authorities. More sophisticated AI tools could help identify patterns across these datasets and flag cases requiring closer examination.
Jigar Doshi, country head - Indirect Tax, Ascentium India, said the department's scrutiny and audit process was no longer limited to information submitted in a particular return.
“Increasingly, the department is able to run data-driven simulations across multiple sources — including Income Tax Returns, GST returns, e-invoices, e-way bills, registrations and historical taxpayer behaviour — to identify anomalies and potential tax risks,” he said.
GST audits go deeper
GST administration is also making greater use of centralised analytics to identify suspicious transactions and taxpayers. Platforms such as DGRAM are used to flag fake registrations, suspicious input tax credit (ITC) chains and potential tax-evasion patterns, while BIFA (Business Intelligence and Fraud Analytics) acts as a risk and exception-reporting engine using the all-India GST database.
At the state level, tax administrations are also developing their own analytical capabilities. Maharashtra's Business Intelligence & Data Warehouse (BIDW), for instance, combines a data warehouse with an analytics layer to help identify cases for scrutiny and audit.
“Manual audits and sample testing alone may not be sufficient against AI-enabled regulatory scrutiny,” says Rajat Mohan, managing partner at AMRG Associates, adding that businesses are facing greater scrutiny of historical transaction data and cross-period discrepancies.
GST audits can involve reconciliation across GSTR-1, GSTR-3B, GSTR-2B, e-way bills and e-invoices. Automated systems can identify differences in ITC, outward supplies and turnover, as well as claims involving suppliers whose registrations were subsequently cancelled or who failed to discharge their tax liabilities.
The scale of information available to tax authorities is significant. India had more than 1.65 crore GST taxpayers as of May 2026, generating millions of invoices and returns that can be analysed alongside income-tax information.
This allows authorities to identify issues such as unusual refund claims, short payment of GST, excessive ITC claims and inconsistencies between sales and purchases.
The growing ability to combine information from GST, income tax, customs and other government systems could allow tax authorities to build a wider financial profile of taxpayers.
According to experts, this means for businesses conventional reconciliation of books with tax returns may no longer be sufficient. Companies may increasingly need to identify potential risks before they are flagged by the department.
AI could also allow the department to alert taxpayers to potential discrepancies, seek clarification digitally and escalate cases for formal scrutiny where required.
However, AI is likely to remain an aid rather than a substitute for human judgment. Officers would still have to verify facts and take responsibility for assessment decisions.
“AI is an enabler,” one tax official said, stressing that the taxpayer or professional taking the final tax position would continue to be accountable. AI cannot be treated as a substitute for human judgment, particularly in complex tax matters.
