Leveraging AI in GST Reconciliation

18 Oct, 2021 admin
Leveraging AI in GST Reconciliation


Introduced in 2017 in India, GST has been a continuous work-in-progress to bring about a balance in the revenue generated from indirect taxes. With constant updates and new changes to simplify the process of paying taxes under GST, the need for a user-friendly, organizational, and management-level application system is becoming more pronounced.

GSTN has introduced several returns that safeguard the interests of the Revenue department and assist businesses in determining the true status of their liabilities with regard to the payment of taxes. Such complexities have given rise to unforeseen issues on the user’s end.

The need for GST Return Reconciliation

A taxpayer under GST is obligated to file certain returns under GST, called GST Returns (GSTR). A major issue faced by taxpayers is the discrepancy in the Input Tax Credit (ITC) claimed by them in the returns. Some of those returns may include:

GSTR 2A – An auto-populated return generated by the system, GSTR 2A shows the inward supplies made by the taxpayer.

GSTR 2B – An auto-populated return generated by the system, GSTR 2B shows the eligible and ineligible ITC available for the taxpayer to use. GSTR 2B is a static return. Once the data is generated for the month, it cannot be amended, even when further adjustments are made.

GSTR 3B – It is a monthly return to be filled by the taxpayer. A simplified return, GSTR 3B, is a summary of inward and outward supplies.

GSTR 1 – A return containing all the details of outward supplies made by the taxpayer. Such supplies include goods and services. All the invoices and debit/credit notes raised shall be contained in this return.

Some issues that may arise while filing these returns are:

  1. Differences in the ITC amount between GSTR 3B and GSTR 2B.
  2. Discrepancies in the ITC claimed under CGST Rule 36(4) and the ITC available to be claimed in GSTR 2B.
  3. The disparity in the ITC claimed in the books of accounts and the ITC available to be claimed in GSTR 2B.
  4. Differences in details of sale between GSTR 1 and GSTR 3B.
  5. Deviation in details of sale between the books of accounts and GSTR 3B.

These discrepancies have serious effects as the taxpayer may lose out on the ITC that should be available to him. As a result, he would have to pay a higher amount of tax in the form of GST.

Another effect of such a discrepancy is that the outward supplies claimed by the supplier in GSTR 1 may not be equivalent to the inward supplies auto-populated in the buyer’s GSTR 2B. This can lead to a loss of credit / excessive credit claimed for the buyer.

Such mismatches in the figures of inward and outward supplies can lead to the incorrect sum of taxes being paid in the form of income tax.

Thus, the need for Artificial Intelligence (AI) systems to stay updated and ease the burden on taxpayers by reconciling these differences is growing.

What is Artificial Intelligence?

AI is a field of science that combines machines, complex datasets, and human expertise to think, behave, and intelligently solve problems.

The need for Artificial Intelligence in GST

With the fast-paced evolution of GST, the pressure to adhere to the changes and updates has increased the workload on companies, accountants and Chartered Accountancy firms. A good application system that provides multi-functional support and replaces manual labor has the following advantages:

  1. It provides a true and fair view of the taxes to be paid and credit that can be claimed by a company.
  2. It can reduce the chances of any discrepancies.
  3. It protects revenue interests by calculating the exact sum of taxes due to them.

An ideal AI application system would provide support by carrying out some of the following functions:

  1. Scanning of purchase and sale invoices and deriving values from them.
  2. Calculation of tax payable and credit available in the invoices of inward and outward supplies.
  3. Segmenting the data available by creating categories for matching documents, missing documents, and mismatched documents.
  4. Providing multiple options to the user to resolve mismatches apparent in the system due to human error, GST return mismatch, and inconsistent document details.
  5. Ensuring that the system is updated with notifications and changes that are introduced and that changes are incorporated automatically into the returns, reports, and summaries generated by it.
  6. Providing regular notifications to the user about the discrepancies and generating messages/emails to the vendors, informing them of such discrepancies, and forwarding replies from them.
  7. Generating various reports and summaries with accurate analysis for further decision-making purposes.
  8. Enabling a detailed dashboard, which would enhance the visibility of ITC discrepancies at both the summary and supplier levels.

An AI application system used by both the supplier and buyer of goods/services will highlight errors within the invoice in the initial stages, thus reducing the chances of a mismatch in the auto-populated returns and returns filed by the taxpayers.

As a result, the amount of ITC credit claimable would be uniform and correct across all the returns, enabling the taxpayer to file correct returns and pay the right tax amount.


With India’s gradual increase in reliance on technology, manual billing systems have become outdated. With the introduction of E-invoicing to reduce GST evasion, the use of such AI application systems has grown. AI doesn’t suffer from human fallibilities and can keep up with the need for accurate, timely, and reliable information. A well-designed AI application system also has the potential to highlight weaknesses within the organization.

Such systems also help relieve pressure on professionals and allow them to focus on management-level tasks. The tasks that take hours to be performed by humans, take mere minutes for AI systems to compute. Since GST is a system almost entirely reliant on technology, it becomes imperative for organizations to invest in AI application systems that are tailor-made for their needs and are aligned with their business objectives.

Whatsapp chat