EY helps clients create long-term value for all stakeholders. Enabled by data and technology, our services and solutions provide trust through assurance and help clients transform, grow and operate.
At EY, our purpose is building a better working world. The insights and services we provide help to create long-term value for clients, people and society, and to build trust in the capital markets.
AI uncovered - Part 2: Decoding tax data: AI's transformative impact
In the third episode of our podcast series ‘GenAI in Tax – The next wave,’ we explore how artificial intelligence (AI) is reshaping the tax landscape. Join EY India Tax Partners Manoj Rathi and Divyesh Lapsiwala as they delve into the transformative power of AI in tax data extraction, transformation, and error management. From reducing manual processes to predictive analysis and dynamic data management, AI is unlocking new levels of efficiency and insight. Tune in as we uncover how AI is not just enhancing tax compliance but shaping the future of tax planning and decision-making.
AI transforms tax data handling by automating extraction, transformation, and error management, thus reducing manual processes and coding complexities.
Predictive analysis powered by AI allows tax teams to forecast trends, partner with business, and enhance strategic decision-making.
AI’s dynamic adaptability enables smarter responses to changing tax data structures, ensuring efficient, future-ready tax compliance.
AI empowers tax professionals to move beyond manual tasks, enabling smarter, predictive analysis and adaptive data management for the future
For your convenience, a full text transcript of this podcast is available on the link below:
Manoj Rathi: Greetings to all our viewers, and welcome to the third episode of our special series ‘GenAI in Tax - The next wave. I am Manoj Rathi, Tax Partner with EY and I am very happy to have Divyesh Lapsiwala, a senior Tax Partner specializing in GST and tax transformation projects, with me today. Together we will dive deep into data, its various attributes, and predictive analysis. So, stay connected as we unravel some of the predictive elements in data today, and how AI will change them.
Divyesh, thank you for joining us.
Divyesh Lapsiwala: It is a pleasure to be here, Manoj. Thank you so much.
Manoj Rathi: When it comes to data, what are some of the most relevant aspects that one needs to keep in mind?
Divyesh Lapsiwala: That is the right way to look at data for any of the tax work we do. The first question really is: where is your data? The second: what do I need to extract and where do I need to get to? Third: during the process, what validations do I need to apply? What are the likely errors, and how am I going to manage them? The fourth is: what is my target format? Where do I need to land all this information? What is the goal of this extraction? And finally, how can I analyze all this information in a more useful manner to be future-ready?
Manoj Rathi: These are great points. But before we talk about AI, I want our viewers to better understand the current ways and the processes that people have put in place around these attributes you mentioned.
Divyesh Lapsiwala: That is really one of the pain points for all taxpayers. The reason is, today, it is a very code-heavy process.
You have a code for every single step of the five steps that we identified, and the code is static. So, it requires the full software development cycle: first, identifying the problem statement, writing the Business Required Documents (BRDs), writing the code, testing, going live, and then realizing that maybe full use cases were not captured, requiring a complete rewiring the project.
If a new use case emerges, you have to rewire the whole project again. So, while today's process is automated, it is a static automation, which creates challenges for taxpayers. It is also heavily code-driven, which means regular taxpayers or non-functional software experts cannot really fix the problem, they can only identify it.
Manoj Rathi: That makes a lot of sense, Divyesh. But when it comes to AI, how does AI change everything in terms of data extraction, data transformation and error management?
Divyesh Lapsiwala: Let us look at data extraction. Today, data extraction today is governed by columnar data that tax uses on a regular basis, and then churns that information from multiple sources. But what if that information is not columnar? What if that information is a mix of text and numbers? How are you going to extract in such cases?
That process becomes heavily complex. But AI’s ability to summarize, index and infer enables you to reach a much deeper connection with all the data sources as compared to what you have today. This eliminates many manual processes. More importantly, if a data source or structure changes, AI’s predictive ability can adapt and suggest the likely throughput for extraction.
When we talk about transformation, it is a similar story. What is transformation? After obtaining the data, you need to format it as required. Now, if you substitute code and use AI, you can feed AI large data sets and train it based on all the transformations done in the past. This helps AI learn and understand what transformations are needed in the future under the same static principles, but also anticipate projected transformations if a new use case emerges. AI knows what you are looking for and makes the right suggestions. Of course, a human will have to make the final call.
Let us take a look at error management. It is a very similar process. But the key point with error management is: can I upstream the error management process? Can I go to the source when the transaction is created? All these learnings pour in and help block an error right from the start. That is really what AI can bring.
If I were to summarize, if I can keep giving feedback, if I can get the taxpayer's team to identify what needs to be done for each new use case, we can possibly eliminate the need coding or recording or iterations that we discussed earlier. AI has the capability to learn in a natural language format, and that is what enables you to bring the power of AI on data.
Manoj Rathi: That is very helpful, Divyesh. This brings me to my last question on AI's ability to perform predictive analysis. In that context, I would like our viewers to understand what is data annotation, and how does AI play a role there?
Divyesh Lapsiwala: That is a new and interesting dynamic that we are bringing to tax. As tax partners or taxpayers, we are not really used to annotations and predictive analysis.
Let us understand what annotation is. It is the ability to tag information from both structured and unstructured sources, bring it all together, and present it for a purpose. Let us consider that I need to analyze of all my past transactions – all my past litigation matters, which could be a combination of text and non-text information – and all my presentations on the risks that I have identified within the organization.
I want to summarize and synthesize all this information to understand what the real risk is, what the risk rating is, what the parameters are, what decisions I have made, and what challenges I am likely to face. Then, I would need to update it for any change in law. With AI, I have the power to take the first steps in this direction.
Today, the entire process is manual and human-loop-driven because now we are not just dealing with data, we are dealing with both text and data. But AI makes this much easier by taking off all of the manual load and leaving the space for intellectual work on the analysis performed.
When we talk about predictive analysis, the story is similar. What is predictive analysis? It involves using complex mathematical models to determine what is going to happen next. How many of us in tax really do this today? We hardly do scenario-building in tax beyond predicting that the rate will move from X percentage to Y percentage, or that the effective tax rate (ETR) will move from one level to another.
But if start using mathematical models on my qualitative tax data, which I have built now over the years, courtesy GST and changes in corporate tax and TDS compliance; if I can start building models for tax, I will not be surprised that tax teams will start partnering with business for making future models of the business for prediction and analysis, and provide inputs and learnings from all tax data through the predictive analysis capability of AI. Manoj Rathi: That is very insightful, Divyesh. Clearly, there is a very powerful story for data in tax. It has been my privilege to host this insightful exploration alongside Divyesh Lapsiwala, and I believe we have uncovered some layers of AI’ potential in tax as we pave the way for a future where decision making is not just responsive, but based on predictive abilities, with intelligence that learns, adapts, and innovates on its own.
As we conclude today's episode, I think we have left our viewers with a very clear view that AI is not the destination; it is a vessel taking you to smarter shores of tax compliance and planning, powered by data. Thanks again, Divyesh, for joining us, and thanks to all our viewers for joining this insightful conversation.
Stay connected with us as we decipher more layers of AI and its power in tax. Thank you and have a nice day.
If you would like to listen to our podcasts on the go:
Stay ahead of the digital tax curve with EY's Tax Technology and Transformation (TTT) network. We help businesses manage tax big data, comply with regulations, and make data-driven decisions.
EY DigiGST solution is one of the largest ASP- GST Suvidha Provider offering a secure cloud platform to meet all Goods and Service Tax (GST) compliance needs of businesses.
