In an exclusive discussion with Shiva from TVW News, Nethrapal M.S, IRS, delves into his 16 years of experience in the Indian Revenue Service, shedding light on the significant advancements in tax enforcement and compliance. He discusses the transformative role of technology, data analytics, and artificial intelligence in shaping a more efficient, transparent, and fair tax system. From simplifying income tax returns to leveraging AI for advanced fraud detection, he shared deep insights into how the tax landscape is evolving to meet the challenges of modern industries.
With over 16 years of experience in the Indian Revenue Service (IRS), you have likely witnessed significant changes in tax enforcement and compliance. How have these changes shaped your approach to ensuring compliance and fairness in the tax system, particularly in light of evolving industry practices?
Over the last 16 years, one change I have seen is that data and technology have become the heart of tax enforcement and compliance. When I joined, my first posting was in a centralized processing center and it was a hub of reengineering and innovation. We were part of a dynamic team that was involved in simplifying the income tax returns to enable their digitization. Originally, the business return ITR-4 was a 24-page long return and then a plethora of changes, including legislative changes, were brought in the board to simplify the business returns. The first major renovation was ITR-1 (Sahaj) and ITR-4S (Sugam), which are the simplest return forms designed to enable simplified return filling, processing, and super-fast refund updates. Today the technology has so advanced and with prefilling, most of these returns are completely prefilled, the processing errors have minimized and most of the taxpayers get their refunds without any visit to the tax office. It’s a big change in the culture itself, where originally the processing itself used to take a couple of years with multiple visits to offices, but the technology enablement has completely transformed the way returns are being filled up to being processed. Over a period of time, the dependence on intermediaries like chartered accountants, especially for simpler income tax returns, has come down substantially. Originally in our times, when I joined, there used to be big tax melas all over India. People just before the due date used to go their queue up and get the return filing done. But today, all these have gone and the end-to-end compliance from return filling to processing is done online with no visit to any offices. In the last few years, even scrutiny and most of the examination processes are completely done in a faceless manner on an online platform, reducing transaction costs and improving the fairness of the tax system.
You’ve designed innovative tools for data analysis and forensic recovery. How do these tools cater to the unique tax challenges faced by specific industries, such as technology, pharmaceuticals, or financial services, and what impact have they had on industry-specific tax compliance?
Yes, part of my work, I was associated with a digital investigation evidence manual and also the commissioning of the new efiling portal for the department. If you look at the digital investigation manual, it lays down a detailed SOP for complex cases on how to take forensic duplication and copying of digital evidence so that the authenticity of the same cannot be questioned in any court of law. The aspects related to handling servers and ERP packages, which are core in large industries, are also done in the manual. Many new forms on how to seize and also collect various digital evidence had been defined. We also came up with many industry-specific case studies on how to do detailed handling of digital data that is present in different industries through different case studies. These case studies have been updated year on year and currently the forensic abilities of the income tax department are on international standards and can be considered one of the best in the world.
However, the next generation of data analytics and big data processing has also been implemented through a comprehensive project that can give end-to-end 360-degree information about any PAN, individual, or company. One area that will emerge is the in-house use of artificial intelligence tools like Knime for data analysis, wherein big data patterns can be found and comprehensive targeting of tax evaders can be done. In-house AI tools and other social media network connectors where network connections can be built can be used to find related parties and associated assets by making various connections. Also linking various databases like GST brings transaction-level data.
In Russia, the entire retail POS moves to the risk management center for analysis. For example, if you buy a mobile today from a shop, the next day you may receive a message on whether they bought this mobile and also check whether the right GST/income tax or anything else has been paid. The future would be completely data-driven tax filing and also system-driven scrutiny of what is being filed. This will enable large processing and scrutiny of various data points that presently sit in various silos. By linking the various silos, we can make definitive patterns and this is where artificial intelligence and regenerative AI may become important.
There are many other areas, like judicial drafting and interpretations, wherein artificial intelligence can play a very big role. Future AI is a tax digital legal assistant wherein a problem statement is given while the AI engine studies all the legal databases, understands, and gives a legal output.
Apart from this, AI will be big in tax fraud detection, where machine learning algorithms can be used to analyze vast amounts of data. Now traditional fraud detection is rule based and some times, fraudsters can build systems against them, but AI can find anomalies and patterns in large data sets that the regular rule-based systems would never be able to find. Using these large AI engines, advanced tax fraud detection would be possible.
The next important thing that we can expect is real-time tax fraud detection. Many a times, real-time tax fraud would enable quicker retrieval of cash and other assets. Now, with big advancements in big data and AI technologies, real-time detection and interception capabilities would be much faster. This would allow immediate action to be taken when suspicious activities are identified, reducing the chances of successful fraud and minimizing losses.
The world is also moving towards advanced predictive analytics; using historical data, we can predict patterns and trends wherein potential risk areas can be identified to implement preventive measures, reducing the likelihood of fraud. With more and more training in AI, the forecasts would become more effective, enhancing the effectiveness of tax fraud detection.
Now one area of innovation on which many countries are working is how to find out manipulated documents furnished to misrepresent financial information. Present technologies like natural language processing (NLP) can be used to verify the authenticity of these documents, understand, interpret complex financial documents, identify inconsistencies, and also identify potential read flags.
Think of this, you receive a mail to upload your accounts in a predetermined XML format downloadable from your accounting software into the system and then the AI engine runs its first set-up of analysis and throws some various analyses and questions that need to be answered. Each and every reason for scrutiny can be system driven and entire scrutiny can be done through digital mode without any human interaction.
When AI gets into collaboration at an international level, then profit shifting becomes extremely difficult. By building global collaborative AI systems, tax authorities can get access to offshore bank accounts and assets wherein new insights can be gained into the wealth evaded. When automatic exchange and verification is done through these AI collaborative engines at an international level, we have a completely different investigation paradigm that will open up.
When we are looking at technology industries, digital tools and AI can be a big deal. Most of these tech companies have sophisticated tax structures in multiple jurisdictions, shifting profits to low tax havens. Since the digital trail is predominant in these, just by analyzing the mailing systems, ERPs, and other document management systems, a pattern can be built on where significant functions and risks are occurring and help the tax department understand the entire value chain in detail.
AI can help build value chains in the companies, which may in turn help allocate profits. We see tech companies billing to Ireland and other tax havens. Such aggressive tax planning schemes can be detected in no time and real-time action can be initiated. Now, unlike the traditional system, the current advanced algorithms can analyze vast datasets to find out unusual patterns and also learn new patterns with new data and can eventually predict and forecast potential risks.
Digital tools are essential in uncovering industry-specific tax evasion schemes. How do you tailor digital investigation methods to address the unique challenges of different industries, such as detecting profit shifting in multinational tech companies or analyzing complex financial instruments in banking?
If we look at technologies that will shape the future tax administration, they are AI, big data, data analytics, blockchain, cloud computing, and the internet of things. For example, an e-way bill transfer of goods can be tracked using the GPS details and complete monitoring can be done in real time.
Now the digital transformation undertaken by tax administration is multiple. Efiling 2.0 has prefilled ITRs with complete AIS/TIS information. There is an enhanced income tax portal user interface that has eased filing of returns and also enabled easier maintenance of e-documentation for all proceedings. Risk management and complete data-driven scrutiny selection by using big data analytics, real-time assessment of risk, and also high-speed processing of returns and issuance of refunds to individuals in particular and complete faceless assessment that has led to efficiency, transparency, and accountability are only the first steps that the Indian income tax department has taken.
The Inland Revenue Authority of Singapore (IRAS) has introduced text mining techniques to analyze unstructured taxpayer email boxes to understand common patterns pointing to common inquiry topics and detect changes in patterns. There can be identification areas of confusion and gaps and also identification of critical mails where tax evasion activity can be detected. For example, searching for words like cash, dollars can yield valuable data.
We will also see invoicing and collection of data at the retail POS more often. For example, let us assume four different mobile phones get sold from four different stores at a particular point in time. We can find out whether GST is paid or not just by looking at the model number and price tag and send out actionable intelligence for further action.
Your publications, such as those on aggressive tax and digital investigation, likely influence industry practices. How have these contributions shaped industry compliance strategies, and what feedback have you received from industry stakeholders regarding their practical application?
Both my publications, Aggressive Tax Planning (published in newsletters of NADT, the training arm of CBDT) and Digital Investigation Manual (published by CBDT) were specifically done to enable departmental officers to tackle the complex world of profit shifting using complicated structures. The digital investigation manual has been widely used in most of the searches and surveys all over India and it is being updated year on year with new methods and techniques. We can boast that this kind of manual is not present with even other agencies like CBI or ED. The income tax department has pioneered the use of advanced technologies and is a pioneer in this area.
On the aggressive tax planning front, the initiation of GAAR is kicking and with time, we will see more GAAR orders out of India.
On the aggressive tax planning side, we have received positive feedback and the work of the department has been widely appreciated. In a recent order on a major retail store, losses that transferred because of a complex demerger process were upheld by the GAAR panel, stating the primary objective was to minimize taxes. With time, more and more landmark cases would be accepted and many industry leaders now consider aggressive tax planning a threat and are redesigning their tax structures to pay minimum taxes in the country.
We can expect more GAAR proceedings, especially where complex merger/demerger and business restructuring are structured to minimize taxes. This is a specialized area and as expertise on various structures gets built, the department will be in a better position to handle these cases.
On the digital investigation front, we are already pioneers and have enabled big usage of the same. The success of many of our investigations is because of the comprehensive use of these technologies.
Advanced analytical tools and AI are becoming increasingly important in tax enforcement. How do you balance the use of technology with industry-specific expertise to ensure accurate and fair assessments, especially in sectors with complex tax structures like oil and gas or telecom?
AI applications are going to be big in tax enforcement. Some likely use cases are
- Autmated detection of anomalies and discrepancies in tax returns.
- Shell companies to claim input tax credit and to create bogus purchases, share application money, and unsecured loans are common. These can be unearthed by building complex networks of these shell companies through a shell company database to uncover massive evasion.
- Risk alerts for high-net-worth individuals and corporations who have high risk of tax evasion to tackle their real-time transactions
- Review of loan documents and contracts to identify pain points and also identify hidden income sources and also verify whether the contract is genuine or not.
- 360-degree profiling of all transactions from various data sources to give one comprehensive taxpayer picture. This would be processed and then prefilled to come up with a tax filing return in an automated way. The data mismatches can be first validated and shared with various intelligence agencies for further action
- Identification of patterns suggesting intent to evade taxes in taxpayer communications. AI can detect communication sensitivity and figure out communications that show a higher probability of tax fraud.
- Prediction of tax evasion likelihood based on past audits and known schemes. AI can come up with a likelihood score of audit risk in the current year based on earlier year audit behavior.
- Automation of risk assessments for refunds and tax credit claims to prevent frauds
- Analysis of legal documents and case laws to support tax litigation and prosecution. AI legal analyzers can be a game changer. Automated drafting of orders may enhance productivity and also increase chances of winning in the appeals stage.
- AI can find tax code loopholes employed by organized tax planners and other gaps and such big tax policy changes to plug tax evasion.
- AI can track the flow of cryptocurrencies to combat evasion through digital assets.
- A key feature that has worked in making taxpayers pay taxes is nudge letters. AI can generate tailored emails and letters that can nudge taxpayers to pay more taxes. For example, making customized letters on payment behavior using historical data and nudging the taxpayer to pay based on personality type may be key enablers of AI.
- Creating a knowledge base of all key learnings, SOPs, and documents from earlier experiences and ensuring the right investigation methods are taken up by the tax investigators to avoid various types of risks.
- AI can help in the development of chatbots to handle routine taxpayer inquiries.
- Offshore data exchange change helps in identification of undeclared foreign asset holdings and income sources.
- AI can help in the automation of report generation for audits, investigations, and litigations.
- Now the AI engine can review emails, messages, and documents and detect intent for tax evasion. Specialized AI software can be designed and companies can be mandated to run these for internal audit controls and reports uploaded to departmental servers.
- Analysis of social media posts and networks can help identify unreported income assets like flats, diamonds, cars, etc. and also help in locating unreported income sources. All social media posts can be tagged on a server and AI engines can be run to get unreported income sources.
- AI biggest delivery would be in figuring how to design algorithms to target limited audit resources at high risk individuals. This would be a game changer.
- AI can also figure out various tax structures and also help in unearthing large tax evasions in Multi national enterprises.
Looking ahead, what is your vision for the future of tax administration in India, particularly in relation to rapidly evolving industries like fintech, renewable energy, and biotechnology? How do you plan to continue driving innovation and efficiency in tax enforcement to meet these challenges?
According to the latest Global Tax Evasion Report 2024, more than $12 trillion, equivalent to 12% of the world GDP, is held offshore by households at the end of 2022. This number includes financial assets only and does not include real assets such as art, gold, yachts, or real estate. One more disturbing pattern is that the offshore wealth owned by low- and middle-income countries like India is increasing compared to high income countries like the US and UK. The tax planning is more shifting to low and middle income countries as per the latest tax evasion report 2024.
However, many feel that there are multiple compliances being done for different agencies, where coherence can be brought in. For example, almost 67% of the corporations in a deloitte tax survey mentioned that income tax compliances can be made better with other compliances like the GST returns, annual return filing with MCA and FEMA reporting. The same set of information is being used at multiple places. So a single point window is something that many taxpayers are asking. However, to do these, the databases that are working in silos should be linked up and some technology platforms are on older legacy systems, while the more modern ones, like the GST and Income Tax Department, which are built on API architecture, can be integrated to other systems easily. Now matching all this information may also be a big plus for the income tax department when it comes to assessing tax evasion risk. The department has already undertaken complex integration of the GST database and data mining and risk assessment have already started. However, broader integration can be expected in recent times. Prefilling of various informations are again happening in the current system. However, we can expect more such integrations in the near future.
One of the proposals made in this survey on income tax digitization in India is that a standard audit file for tax (SAF-T) may be introduced, which is a global OECD standard and has been adopted by many European countries. This benefits taxpayers since their process of submission of data is completely automated and also significantly enhance and automates the tax audit process. The survey states that once this SAF-T is introduced, audit of both direct and indirect taxes can be completely centralised and independent examination through data analytics or AI can be carried out to enhance better audit outcomes and risk assessment for the department.
A lot of suggestions are made on how reconciliation can be done with their existing ERP systems. For example, if TDS credits are linked to GST invoices, which again are captured in Form 26AS/AIS, then it would be easier at the corporate level to track and claim them in the income tax return and eventually help in reconciliation.
One more significant area where technology has to be used as per this survey is to track the status of refunds and where it is, appeals disposals and giving effect to the appellate orders. These can be expected in the coming days, where large corporations can expect timely integration of pending disposals and likely time to dispose in a single dashboard.
Note: We are also on WhatsApp, LinkedIn, and YouTube, to get the latest news updates, Join our Channels. WhatsApp– Click here, and for LinkedIn– Click Here