Future of AI & ML in TaxTech space

One of the most critical challenges in the tax space is keeping up with compliance. Tax compliance is a dynamic labyrinth of complications, and it can be overwhelming to navigate. That is where TaxTech comes into play. Essentially, tax is data. But its volume is not something that you can handle manually. If you do attempt to handle tax compliance manually, the scope for human error will be high. The time and effort to manage tax compliance is very high. Artificial intelligence and machine learning solutions can help automate tax compliance for organizations irrespective of the size or nature of their business. Today, AI and ML are being used to sift through massive amounts of structured and unstructured content to extract tax rates, rules, and jurisdictions. TaxTech solutions can help your business assign the right tax codes and tariff codes to the right products and ensure they meet tax compliance requirements anywhere in the world. Document verification solutions use AI and ML to review and validate a variety of compliance documents. AI and ML are essentially providing the path for businesses to scale without worrying about tax compliance violations. This is not about reducing manual labour anymore. AI and ML are helping automate tax compliance space in a way no one could have thought possible.

Well, there are business as usual functional approaches while using AI and ML. These include assigning the right tax codes to various products so that they can be taxed correctly at their destination. Let me give you an instance. A business sourcing local products from a country will need to ensure the correct tax and tariff codes (based on the nature of the product, its price, and the destination state or country) are applied to the products so that they can be accurately taxed upon arrival at the country of delivery. While this might sound simple enough for one product, imagine repeating this exercise for 10,000 different products in many categories.

By assigning the right tax codes, keeping up with changing tax policies and auto preparing and filing tax returns, organizations can maintain transparency. All the relevant information regarding the product and the consignment can be easily generated and transferred to a QR code for easy access later. This will allow revenue authorities to know everything about your consignments easily. It reduces the friction between organizations and revenue checkpoints. There are fewer delays at customs. Additionally, the fact that you can maintain transparency, allows authorities to differentiate fraudulent businesses from ones that are doing an honest day’s work.

For a customer, a transaction is essentially making a payment for a purchase. They are likely to focus on the product, its price, shipping costs and taxes. But there is a lot more to do for the seller. With the help of technology, it becomes easier to automate various processes such as payment collection, localising the currency, calculating shipping, taxes, tariffs, custom duty costs etc. With AI and ML systems, the technology can learn about how those processes are being executed and develop algorithms to better automate those processes, eventually leading to autonomous compliance, or compliance that does not require manual intervention.

There are 3 steps to locate the right tax data. First, is digital documents. Technology such as Optical Character Recognition (OCR) and Natural language Processing (NLP) can take a physical document and digitize it and standardize the data in the document. The second is the cloud. Putting the digital documents in the cloud makes it possible to access them from anywhere as well as to have them all in one location. Finally, AI and ML makes it possible to understand the data in the documents in the cloud so you can find what you need, accurately, from wherever you are. This approach can be applied to managing tax content as well. The United States alone has over 15,000 tax jurisdictions. Each tax jurisdiction can have its own tax policies, tax rates, tax deadlines, even tax holidays. It can be extremely overwhelming to have to deal with so much data that changes quite often. AI and Ml can automate the sourcing of this data from all the right places, structuring it so the system can understand what has changed and by how much, store it in the cloud and then use it to find the right tax rate, for the right product, for the right location. Not only are these systems helping you work faster and better, they are essentially helping you save on any additional expenses arising from missed tax liabilities.

The destination is autonomous compliance. ‘Self-driving’ compliance in other words, where digital documents and data power automated processes and workflow that take advantage of machine learning to make decisions and deliver outcomes that have previously required human judgment. Not every process and workflow should be autonomous. There still needs to be human oversight for training, verification and control, but it will be possible to automate almost all of the steps in compliance to get the best of technology and human expertise.

Author: Sanjay Parthasarathy, Chief Product Officer, Avalara.

Publication: SiliconIndia.