The Federal Board of Revenue (FBR) has decided to monitor over 7 million income tax returns using Artificial Intelligence (AI). According to sources, the total number of income tax returns during the current fiscal year may reach up to 8 million.
Tax returns reflecting a tax payment 25% lower than either FBR’s assessed estimate or last year’s return will be subject to audit. This large-scale audit will involve 8,000 Inland Revenue officers and around 4,000 auditors. Major companies, various sectors, and individual taxpayers will be included in the audit.
FBR plans to process 7–8 million returns through an automated and digital AI-based system. Initially, Microsoft Excel filters and formulas will help flag potential cases, while AI and other digital tools will assist in deeper audit analysis.
Companies or sectors that have paid at least 25% more tax compared to the previous year will automatically be excluded from the audit list, whereas those paying less will be prioritized. All submitted returns will undergo technological screening, and businesses paying lower taxes than the previous year will be the primary focus of the audit.
The audit will be conducted in three phases:
- Commercial and industrial income tax returns will be reviewed first.
- Single-owner and other registered companies will be audited in the second phase.
- Individual and salaried taxpayers will be covered in the final phase.
According to sources, taxpayers found to have provided false information, displayed disproportionate wealth on social media, hidden assets, or held benami (undeclared) properties will face action.
The audit will examine tax returns from eight different analytical perspectives. By October 31, 2025, around 5.9 million returns had already been filed, with individual taxpayers contributing around Rs9 billion more in taxes compared to last year.
The deadline for manual tax filers has been extended to November 30, 2025, by which time the total number of returns is expected to reach 7 million.
